Oxide and Friends

Oxide Founder and CEO, Steve Tuck, joined Bryan, Adam, and Oxide Friend, Steve Klabnik, to talk about our recent announcements: general availability of the Oxide Cloud Computer, and raising $44m. The reception was (broadly) great! Bryan and Steve answered questions about the product, company, and launch.

In addition to Bryan Cantrill and Adam Leventhal, we were joined by Steve Tuck and Steve Klabnik.

If we got something wrong or missed something, please file a PR! Our next show will likely be on Monday at 5p Pacific Time on our Discord server; stay tuned to our Mastodon feeds for details, or subscribe to this calendar. We'd love to have you join us, as we always love to hear from new speakers!

Creators & Guests

Host
Adam Leventhal
Host
Bryan Cantrill

What is Oxide and Friends?

Oxide hosts a weekly Discord show where we discuss a wide range of topics: computer history, startups, Oxide hardware bringup, and other topics du jour. These are the recordings in podcast form.
Join us live (usually Mondays at 5pm PT) https://discord.gg/gcQxNHAKCB
Subscribe to our calendar: https://sesh.fyi/api/calendar/v2/iMdFbuFRupMwuTiwvXswNU.ics

Speaker 1:

We get some comments from time to time about

Speaker 2:

Me too.

Speaker 1:

And I hope you're sitting down about us being lousy podcast hosts.

Speaker 2:

We are. Well, I mean, I don't understand how anyone can criticize our intro music for starters.

Speaker 1:

That's right. There's very little to critique.

Speaker 2:

I mean, it is, it is beyond reproach. Yes. It it may be true that our intros are a little terse.

Speaker 1:

Terse.

Speaker 2:

Also non existent.

Speaker 1:

Well, there's that. But today, I think particularly one of the one of the reasons we get dinged is by not introducing our guests. So, especially because we have 2 Steves as our guest. I thought we should introduce our guests. So Steve Klabdick is Steve Klabdick is here with us.

Speaker 1:

Steve, welcome.

Speaker 3:

Hello. Hello. Thanks.

Speaker 1:

Friend of the show, oxide colleague, and I don't know, generally famous. Is that fair, Steve?

Speaker 2:

No. Just generally famous. Did you would some dude.

Speaker 3:

I I add them to Discord, though. That's a good that's a good, credential.

Speaker 1:

But number 23? 23.

Speaker 2:

Most I hacked the karma all time.

Speaker 1:

All time on the planet. Yep.

Speaker 2:

He is part of the, I dare say, the the hacker news aristocracy. Yeah. He's a baron on hacker news.

Speaker 1:

How do you how do you spend all that karma?

Speaker 2:

Yeah. How do you well, in this is I am convinced that there is, like, a global services for Hacker News that they can't. I think that he hits, like, he hits, like, 30,000 and, like, something different happens. And he's just, like, we're never gonna get it out of him. This should be, like, a little twinkle in the eye.

Speaker 2:

I'd be like, no. It's just like, I don't know.

Speaker 3:

Yeah. I I told myself I was quitting a 1,000 points, and I told myself I was quitting at 10,000 points, and then here I am forever later. But I I feel like this is not in service of my cause, but I wanna be like, actually, I'm number 24 because Paul Graham is hidden off the leaderboards, and he has the most. So if

Speaker 1:

there's one on top,

Speaker 3:

but, yeah, I don't I don't feel like I was defending myself very well.

Speaker 1:

Right. But who's counting? Right?

Speaker 3:

Yeah, exactly. No one's

Speaker 1:

counting. And, other Steve. Welcome to other Steve. Steve Tuck.

Speaker 2:

Oh, was this meant to address the criticism that our intros were bad? Is that I mean, I'm just trying to understand. Are we trying to okay.

Speaker 1:

No. I'm saying, look. You thought that was bad? Could have been worse.

Speaker 4:

That's right.

Speaker 2:

Like, okay. Okay. We'll do it your way, pal. We will do intros. You'll be begging us to never do another intro again.

Speaker 1:

Exactly.

Speaker 3:

Thank you.

Speaker 4:

That was very kind.

Speaker 2:

Yeah. Other Steve,

Speaker 1:

other Steve, I guess, our, our CEO, and oxide cofounder, of course.

Speaker 4:

Bottom of 25 karma on news. Yeah. Hold on. My my karma my my karma also was hidden.

Speaker 1:

That's right. Yet to earn karma on hacker news, but any day now. So big week last

Speaker 2:

week. It was big. Yeah. It was, it was. And, a little bit surprising, I guess.

Speaker 2:

I mean, I know that's the wrong thing to say. It's it's like our launch. We shouldn't be surprised, but, I was a little surprised. I was surprised. I was just flat out surprised.

Speaker 1:

I'm surprised by how surprised you are just cause the, the Internet's love us.

Speaker 2:

That surprises me.

Speaker 1:

There you go. You're just easily surprised, the Internet's love. So we, are we, we announced the the first shipment, or the shipment of the first cloud computer. We announced the raise, which is, you know, kind of a big deal because it pays all our salaries. And, and everyone loved it.

Speaker 1:

Like, the the Internet's lit up, and people are very interested. And I was unsurprised by their interest.

Speaker 2:

Yeah. It it was great. I mean, it was it was honest it was delightful, honestly. And I guess it's also there's a little bit okay. Now this is, is like let me just get this get this out there.

Speaker 2:

Let me just get this this odd off my chest. There's a little bit like, why is everyone so surprised that we shipped? You know what I mean? So what it was like you know, it's like yeah. I kinda feel like, you know, it's like when I graduate from college or whatever.

Speaker 2:

It's like, people don't need to be that surprised. You know? You can actually, like, act like you knew this was gonna it was gonna happen this way. But I think it is I mean, but, of course, truthfully, it is surprising, and they got every everyone's got every right to be surprised.

Speaker 4:

I mean, we also talked about the entire journey. We did. The near misses, the difficult times.

Speaker 2:

Well, I which is the other thing that I did think that so okay. Adam, this is the other element of my surprise is that and and we did there were some stories and journalists were like, this company is coming out of deep stealth. It's like, we are aren't we the ultimate?

Speaker 1:

Oh my goodness. Like, how

Speaker 4:

Much like stealth TV.

Speaker 1:

We get out here and proverbially, to be clear, open the kimono, like, every week. We're talking about some esoteric detail of the product. I mean, some problem that nobody's ever talked about, but everyone has had. I mean, I don't know what deep self is. I don't know how we can be more the antithesis of that.

Speaker 2:

Yeah. If anything, we're like TMI. It's like, hey. Let's just I'm up already. Do you guys, like, have some problem that you just, like, keep to yourselves for once?

Speaker 2:

Like, experiment with that. Like, I just don't need to know everything. It's, like, kinda gross, honestly. Go to therapy. It is it is therapy.

Speaker 2:

There's truth to that.

Speaker 1:

But the podcast is cheaper than therapy.

Speaker 2:

The podcast is really cheaper than therapy. And the so it would that part was I mean, it was and, obviously, there are plenty of folks online who were like, these it's kinda been talking about what they're doing for a while. But I think it's also easy for us to kinda get in our bubble and realize that, like, not everyone listens to the podcast, and not everyone is following what we're doing. So, everybody's on Hacker News. Not everyone's on Hacker News.

Speaker 2:

I'd, I it's as I try to remind myself, or or Yacker News as my kids call it. I'm doing yacker news again. It's like Burn. Come to your room.

Speaker 4:

Lobsters as I came to find out.

Speaker 2:

Or lobsters. What that we we need to came to find out. What does that mean?

Speaker 4:

I just wasn't aware of of lobsters. I mean, aware of, but not I I mean, I had think I'd been pointed towards it once or twice before.

Speaker 2:

Well, and were you I mean, as long as we're talking about all places in which we are discussed, were you aware of slash dot?

Speaker 4:

I mean, yeah, that's

Speaker 2:

Oh, man. That was, that was definite one step further. Yeah. I I would say never change slash dot, but they're one step ahead of us in that regard. So, so yeah.

Speaker 2:

I mean, I I don't think it might be interesting to to, go a little bit kind of into the the mechanics of the launch. I don't know if people are are interested in this, but just in terms of, like, what's actually involved in a you know, what it means to actually engage journalists, for example. Because there were a bunch of news stories that that, came out, that we were all kinda coordinated, and it's like, how does that actually happen?

Speaker 1:

Yeah. So so how long ago did it start? Like, what when did you start having conversations with PR and and journalists, and and how did it roll out?

Speaker 2:

Yeah. So I mean, I think we so we started probably in earnest maybe, like,

Speaker 4:

2 months ago. Probably, like, yeah, several weeks ago.

Speaker 2:

And we, knew that, we you know, what one of the advantage I think we've said this we've certainly implied it over and over again, but just to say it explicitly. There are a lot of advantages to starting a company later in your career. And, Steve, I think you and I have had have just really appreciated that. And then we've got because we've got a team of veterans that are all I mean, we're not, you know, been around the industry for a long time. And so everybody's got a really big Rolodex.

Speaker 2:

So when it came to, like, a PR firm, well, we we we know, like, who's on that shortlist and, you know, ran a process and, which something we emphatically believe in. It's like we don't wanna be we always wanna run a process for this kind of stuff, and you're kinda selecting a firm like that. And, it was it I feel like this has happened, certainly more often than not, Steve, where we kinda come in with our but we've worked really well with this firm. Let's run our process and and see what happens. And then it's like, you know what?

Speaker 2:

The firm that we love is actually the best firm. So that was actually great.

Speaker 4:

Because, I mean, it it's pretty easy to fall, and I think for a lot of companies, to fall into the playbook that is out there, be it a playbook that a set of investors will put in front of portfolio companies.

Speaker 2:

For sure. Yeah.

Speaker 4:

Folks in the network. And once people have gone down a certain path, they're gonna kinda recommend it In not necessarily if they've had a bad experience, but you will you will see these kinda just add water playbooks where it's like, you know, insert mega PR firm over here or law firm or

Speaker 2:

I mean, it's like when you get these kind of because you you need these professionals.

Speaker 4:

Full payroll platform here. Sure. And and PR is just another one. And I think there are a lot of companies that aren't sure if they need it or not, and they are told, like, no. No.

Speaker 4:

No. You definitely need this. And it can be in cases where they haven't even shipped a product. They haven't, like, started on the product. And there's a lot of lot of ways to spend money poorly.

Speaker 2:

Yeah. And I think that, you know, PR actually, Steve, you know, I haven't talked about this. It'd be interesting to get your take on it. I think PR is one of those things, like sales, that's often misunderstood by technologists. They think, like, oh, the way I get people to care about my thing is PR.

Speaker 2:

And the way I I get my thing sold is sales. And it's, like, well, that's that's that's part of it. But it's, like, PR can't make people I it a a PR engagement or engaging with a journalist can't make people care about something that people don't care about.

Speaker 4:

And No. Worse than that. I mean, you now are, like, who bridges away from the actual message? Right. So if you don't have the right firm who's talking to over clergymen, like,

Speaker 2:

you know, you are report

Speaker 4:

more diligence than others. And, and so you're already gonna get a version of the story. Right? Not your version of the story. And then, you know, if you have the wrong folks in the middle, it can go really crazy,

Speaker 2:

I think. Well, and in particular, you can have, like, a bad PR firm. It could be like, you know what? Actually, like, your story is actually pretty uninteresting, so I told a different story to this reporter.

Speaker 4:

They love it.

Speaker 2:

They love it. They are super interested. Like, okay. Well, that's also interesting. That's a big one.

Speaker 2:

Yeah. That's not what we do. So, you really need a firm that kinda understands who you are as a company, kinda the soul of the company. And, we're really lucky with the Long Squad folks. Definitely have worked with them in previous lives, and and they've been really terrific.

Speaker 2:

And in particular, one one of the things that part of this, I I wanna call them out in particular because as they were kind of, they under absolutely understood Oxide. I love the fact that they'd come in listening having listened to the podcasts, and they you know, that's always an a metric for us is, like, have you done your homework on us? Do you understand us? Of course, they had. But in particular, they're like, you know, we really see the transformative power of this, and we think strongly that it needs a name.

Speaker 2:

That we need to like, not a product name. It needs a category name. And we agree. You know, I think it's and the, you know, and I think it was it was, you know, their suggestion that, you know, we think you should call this the cloud computer. And does that make sense?

Speaker 2:

And, you know, that and we were kind of trying it on. I, you know, one of the things I whenever you get something like this, you wanna you wanna test that out a little bit, And, Steve, I remember, like, the the conveyor suggested it in the morning, and I think we had, you know, 2 different customer meetings over the course potential customer meetings over the course of the day. And we were both kinda trying it on, and, you know, watching the reaction to it, and it was helping to catalyze it. It was really helping people understand, because there's a kind of an interesting built in juxtaposition with juxtaposition with cloud computer, because we we kinda think of the cloud as being something that we don't see that that that we are, that is often the public cloud, and to have this kind of juxtaposition of, like, no. This is this physical thing that runs Elastic infrastructure.

Speaker 4:

Cloud cloud computing is running in someone else's data centers. Cloud computing is something you rent. Cloud computing is this kind of ethereal service layer that you can't see, can't touch. And, in fact, go into the data centers of where where these services are running, and there's, very quickly understand that there's a very the big difference in what's been built to support that. Totally.

Speaker 4:

And so we kind

Speaker 2:

of view this as, like, just like you've got a microcomputer and a mini computer and a personal computer and a and a mainframe computer with this is the cloud computer. So, that was something that was, I thought that was a a great suggestion on their part, and as the more we were test driving it, the more made sense to us. So, as we were, you know, kind of the way these things work is that you, and this is you kinda go pitch, reporters on a story. And the thing you have to understand, and I I I think, I'm gonna speak up on behalf of the 4th estate here, because I really think that we have culturally really maligned journalism, and especially in Silicon Valley. There's this idea that journalists are the enemy.

Speaker 2:

And, you know, my wife's a journalist, and I I know a lot of journalists, and I I really take issue with that kind of perception. Of course, there's bad journalism, just like there's bad everything. I mean, there's there are in any profession, you can find the worst of the profession and the best of the profession. But the best of journalism is important, and terrific, and has broken huge stories. So we that that's all the the things in the positive category of journalism.

Speaker 2:

One of the challenges with journalism is they are telling a story, and, you know, when you are, I mean, we think that there's a big story here in the in the abstract, but it's a big story. It's not one that's it it it's not something that is a Doesn't

Speaker 4:

fit in 500 words on them.

Speaker 2:

And it's not like, what's the news? You know, the there's a difference between us having a story and having something that's, like, newsworthy. And so you gotta, you know, couple, the the larger story, with something that is newsworthy. For us, I mean, there are kinda 2 newsworthy elements. Obviously, there's that that we are, that we are in customer data centers right now, and we are announcing our general availability, and then we're also announcing our fundraising, which, was actually wrapped up a while ago, but it's just not something we've talked about.

Speaker 2:

So this was gonna be our opportunity to talk about that. We were we're we wanted to be very circumspect about the way we talked about fundraising. I mean, Steve, you and I have both seen, and Adam, you too have seen we've seen plenty of companies that treat fundraising as if it's the objective. And it's like fundraising is not the objective. Fundraising is something that is necessary on the way to the objective.

Speaker 2:

So we wanna you always wanna be careful about kinda how you do that. But we do wanna we also need to announce it because we had I mean, we hadn't really said anything. And, you know, on the one hand, our investors like to keep a low profile. On the other hand, like, okay. Like, not that low profile.

Speaker 2:

Like, can we can we start talking about this?

Speaker 1:

So it's it's a go hand in glove too to to say that we've got this product in market. It's available now, and we're not some underfunded startup. We've got this war chest. So so don't worry. We we're gonna be able to support this thing and actually bring it to market in the way that your business is needed.

Speaker 2:

That's right. Yeah. That's right. And I think it you need to know kind of how I mean, that should be important to, and I hopefully, that is important. But, yeah, Adam, that's a very good point, and that that war chest really is, it is essential to know that, like, okay.

Speaker 2:

I'm I I I am gonna be you know, this thing is early, but there's gonna be longevity here. And there are a bunch of investors who really believe in this thing, and they include the very best of institutional capital, institutional venture capital, and some of the very best of corporate venture capital. And there's some just really big backers here, which is terrific. And and we obviously really appreciate all those folks. So, yeah.

Speaker 2:

They Adam, you're right. I mean, it's a it is a really it's an it is an important hook. But so as we were kinda thinking about, like, alright, so we need to go, structure this, want then you get to the kind of, like, outreach to journalists, and you're trying to tell them what the story is. And of course, like, and out of my camera, we you know, I talked about this, but it just sounds a lot like talking to venture capitalists in terms of, like, oh, like, we're too early for them or we don't do hardware. I'm like, this sounds like you're talking to VC guys.

Speaker 2:

Like,

Speaker 4:

where the raise is not above a certain level. So right. Exactly. You're

Speaker 2:

like, okay.

Speaker 4:

Alright. So I'll be writing from the sidelines.

Speaker 2:

There was some of that, actually. Yeah. Exactly. And and, actually, to their credit, I mean, we've known a lot of journalists over the years. I mean, Adam, I do think it's kind of funny that we knew Ashley Vance, like Oh, yeah.

Speaker 4:

When

Speaker 2:

he was a little too proud. And And we I mean, we were proud. We were all so proud together.

Speaker 1:

I think was that article from, like, 2004, 2005? 2004.

Speaker 2:

Yeah. 2004. And that is the so the the for the the so the very first article ever written by any technology that I'd had any part in was the article that the register wrote on DTrace and featured a profile of the 3 of us, of Mike, Adam, and May. And Adam, the 3 of us sitting in my apartment because we didn't have a, I, didn't have a better photo of us holding a, scrabble rack as I recall. That one had a bit of a surprise.

Speaker 2:

I know. When

Speaker 1:

you say it out loud, it sounds

Speaker 2:

weird. Was Ashley did Ashley physically take that photo in No.

Speaker 1:

No. No. No. Remember, you and I went out drinking with Ashley. And, in I think it was like Sun Network in China.

Speaker 1:

I know I'm saying this and it sounds like a dream.

Speaker 2:

It yes. Go on. And then what happened?

Speaker 1:

And and then Was Steve there?

Speaker 2:

Or was it or was I there?

Speaker 1:

You were there. There were no Steve's there. And then, the headline was something like some engineers drunk on something or other. And and we looked at it and said, oh, what is gonna this gonna be? But it turned out to be pretty good.

Speaker 2:

So I it so in my dream, Ashley took that photo himself. Or maybe not. So you think Bridget took that photo? You think my wife?

Speaker 1:

I think so. I think that's right. Yeah. I think that I mean, I have it in my in my photo album, so I think that's domestically generated.

Speaker 2:

Right. So anyway, the, it and I but actually went on to so left the register on went to, I think, at the Bloomberg New York Times, one of those stops, and Was it the he was at the New York

Speaker 1:

Times. Yeah.

Speaker 2:

Yeah. And it was at the New York Times and then wrote, what is one of the the kind of the the important books about Elon Musk and is now, like, a traveling, like, tell he's a television host now of a television show, which is great. So Ashley has become and, I mean, on it it I mean, the piece he wrote was really good, actually. He, went deep on it, and he, real he talked to our users, which was great. So, but so on the one hand, we've cultivated

Speaker 4:

That's that's what good looks like.

Speaker 2:

That is what good looks like. Is like That is what good looks like.

Speaker 4:

Yeah. Reporter that actually wants to dig deep, understand the people behind it.

Speaker 2:

Understand the people behind it.

Speaker 4:

Like, the first couple people, but, like, the team and and where things have been in the industry for a per period of time, why this is new, different, etcetera. That's what you're after.

Speaker 2:

Yeah. And and I you know, the the fact that he likened me to Ron Papale, I I think I this is have you ever seen, like Oh, yeah.

Speaker 1:

I I remember you saying that your mom and sister, like, dined out on that for, like, months.

Speaker 2:

Yeah. Dined out on my carcass. They dined out on me. Because and is this in the story or not? I okay.

Speaker 2:

So I Ron Pyle was an infomercial Yes. Salesman. Yes. I And

Speaker 4:

I just had to see his picture.

Speaker 2:

I I and I know him you know, I'd like to say I I would like to hope I'm in a safe space. Yeah. Thank you. Someone has dropped the picture to the chat just in case. There's, like, how can this be so bad?

Speaker 2:

Oh, goodness gracious. Okay. So he was a a famous pitchman, and in particular, I I had purchased a food dehydrator. I know this is not a safe space. Did you do not know this, Steve?

Speaker 2:

No. Yeah. And Adam, you may have forgotten this. Like, why am

Speaker 1:

I No. I'm I'm I'm I'm here for this. I'm

Speaker 2:

I'm I'm

Speaker 1:

I know this one.

Speaker 2:

This one. Is the one that you've left to your heirs to be opened upon. That's

Speaker 1:

right. My untimely demise. Yes.

Speaker 2:

Untimely demise, especially if if pride does not have an airtight. Oh, I the, so I had bought a food dehydrator. You'd be very surprised to learn that they're basically garbage. So this is what because so you watch this info merchant, god, this seems great. And Problem number 1.

Speaker 2:

Okay. And so it's basically just like a a cheap piece of plastic that that heats that that at extremely low temperature dehydrates food. And so I'm like, I'm gonna make raisins. So I go to the grocery store, and this is just embarrassing, like how this is the part that they don't feature on the infomercial that you go to weigh the raisins. And I'm like, wow, that's $12 worth of raisins.

Speaker 2:

And of course, I'm like, I have just paid for the water in the grapes. I bought the grapes, and it is the water that I have paid for, and I'm now going to go dehydrate these grapes and turn them into raisins. The I would like to say that I made the world's most expensive best raisins. And yes, if you're wondering, am I the right are we just for the oxide launch? Should we talk about raisins?

Speaker 2:

We're talking about both. They're related. Just wait for it.

Speaker 3:

Startup number 6,040 2 is, finally an integrated, well designed top to bottom, of dehydrator.

Speaker 2:

Food dehydrator. Truthfully, the bigger problem with dehydrator is that you that ultimately, you're paying for the water. When you go

Speaker 3:

Only near that scale, it would be everything would be better, but they're just not. No.

Speaker 2:

That's right. So, anyway, so so when

Speaker 4:

you deploy a traditional rack, what is the net you're left with at the end? That too abrupt of a transition?

Speaker 2:

Too abrupt. So so the, but so Ashley had written this great story about us. And this is, like, on the one hand, we we've, like, got this great rolodex over the years. But if you on on the other hand, like, Ashley is now world famous, and a bunch of the other journalists that we've worked with are just like 2 either too prestigious or retired or both. So it's like, well, alright.

Speaker 2:

That's not that's not very helpful. So we

Speaker 3:

Sorry all your friends are too good, Brian. That's very

Speaker 2:

It sounds very rough. It is. It's very rough. I'm glad that you understand my my dilemma, Steve, about how difficult it is to be. So it it but it is it was tough.

Speaker 2:

I mean, honestly. And so we, as we were, you know, trying to find folks and the the the end launch, we did a terrific job reaching out to a bunch of folks. Maybe we you get to the point where, like, alright. So we got a, you you kinda try to sell an exclusive, but that's that that can be a hard sell when, even though we would love someone to really dig deep. But, anyway, you do what's called an embargo.

Speaker 2:

So you say, like, alright. So we've got this news, and the news is under embargo. And you got a press release that you write, and the press release is gonna go out on the wire, and there's a a couple of different wire services, PR, that there's PR newswires. There are a bunch of them, where you're basically pushing your press release onto a press release newswire, and now that is available for anyone to a story on. And it is a press release is, like, for the press.

Speaker 2:

It's so it's

Speaker 4:

been a press release. Is pretty well recognized.

Speaker 2:

Yeah. For sure. I mean,

Speaker 4:

they don't have to live by it, but most journalists will adhere to it, and it is effectively just so someone is not preempting the news.

Speaker 2:

That's right. And so when you have somebody under embargo, yeah, they're always gonna. I mean, that's just like I mean, yeah, I guess they don't legally have to, but it's but, a reporter is basically gonna honor an embargo, and the so we we kinda knew that was happening. Then once you kinda commit to that date, it's like, alright. You're basically locked in for the date, and that's gonna be and we knew we were gonna write a blog entry that kind of accompanied the

Speaker 4:

the press release. And setting that was about 2 weeks before. Yep. Like, starting to give reporters the opportunity to write under embargo. Reporters will then reach out and wanna do interviews.

Speaker 4:

Right.

Speaker 2:

Yep.

Speaker 4:

Have a bunch of q and a back and forth. But once we had set that, the timing of that was a bit early.

Speaker 1:

Yeah. So why is it at 3 AM? Why where where's that number come from?

Speaker 2:

That that that is 6 AM EST.

Speaker 4:

So and that is just Why is 6 AM EST?

Speaker 2:

Well, that's just so they can

Speaker 1:

Even that seems early.

Speaker 4:

Well, so the the particular reason for that, Long Squad explained, was that a bunch of financial service newsletters go out at 6 AM EST. So if you wanna be included with these, like, financial trade rags that go out, you wanna be launching it at 6 AM east eastern, which I did not know.

Speaker 2:

You well, just in general, if you want a reporter to have a story that's gonna publish on that morning, you wanna give them you don't wanna have, like, a 9 AM embargo where they have to be, like, well, shit. Like, the story's gonna go live at 8:30, and that would be jumping the embargo by half an hour. Like, you just wanna give them, like, alright. So it's gonna be so 6 AM, 3 AM on, here on on the West Coast, but we don't necessarily need to be up for it. So, we kinda have that locked and loaded, and then we we know we need to write a blog entry for it.

Speaker 2:

And so we we're kind of noodling that. It was very catalytic to how we were at a an event that actually, one of our investors was hosting, and they, in particular, a counterpart, they were having this conference, and they you know what would be great is if you had the oxide rack

Speaker 4:

at this conference. Where was it being held? In San Francisco. No. But, like, what kind of a building was this?

Speaker 2:

I just

Speaker 1:

asked Was it just an office building? Or what kind of was it a stadium?

Speaker 4:

I mean, it it was like over 1.

Speaker 3:

Like, what was the zoning

Speaker 2:

of the thing that was on? Yeah. He tried to knock knock me out here. I mean, it's like

Speaker 4:

it's an illegal establishment. It was it was a church, and it was a pretty tricky pretty tricky logistics endeavor to get the thing in there.

Speaker 2:

And we and to be clear with, we did not plug the rack in. And it did we took what we call the showcase rack, which got mechanical blanks and is a little bit, like, less weight. But, That

Speaker 4:

turned out well.

Speaker 2:

It was great. I was I mean, I think we're all, like, a little bit skeptical. It's like, what do you we wanna have a whole rack there? It's not small. It's big.

Speaker 2:

And it's gonna go

Speaker 4:

to a church, and it's like the the well, in particular so this this particular investor, they work very closely with CBC's corporate venture capital teams. And so the folks that attend their annual investor conference are almost all Fortune 500 companies, which is definitely a demographic that is good to have exposure to the product that we're building. And so they they they made the the the case very, very easily. And getting it in and out was a little tricky, but, and I don't think we will find the rack in as luxurious a location ever again. Yeah.

Speaker 4:

No. It was it was it was impressive.

Speaker 2:

It was impressive. Kind of, it was, like, surrounded by, like, tapestries or you know, it was very, Velvet. And it gets you that that, like, that afternoon, like, oh, and it was and it there was a there was a little bit of a some of the holiness around it, I gotta say. A little bit of holy rack. But so at and the kinda the reason I bring all this up is because at the event, there was you know, at these events, you get a lot of, like, very buttoned down to kinda corporate venture capitalists, and then there was always a couple, couple of wild cards in the deck.

Speaker 2:

And one of the the the wild cards was someone who's run a con a consumer brand that you've heard of, and, but was that was, a bit flamboyant the right word? So if if if there were one person at this conference Centric. Dressed to go to a club, it was this guy.

Speaker 4:

In in every way imaginable.

Speaker 2:

In every way imaginable. Like, if you put that in your head, you basically have the right visual. And so, like, a little bit of a Extraordinarily caffeinated. Extraordinarily caffeinated. Maybe generous.

Speaker 2:

Like, not, I mean, at least caffeinated. Caffeinated plus plus.

Speaker 4:

That's right.

Speaker 2:

And so I was, like, kinda waiting to for food or whatever. And, and I ended up basically talking to this guy, and I'm thinking like, man, this guy is not there's just he's gonna he's a he's a direct to consumer person who looks like he just got kicked out of the club. I mean, it's like this is not gonna be and, it turns out, like, do not judge a book by its cover. Because I kinda got half a sentence into what Oxide's doing. And he's like, god.

Speaker 2:

Like, oh, man. I got so many memories dealing with necks and dealing with this and the other. Cables are everywhere. IMU. You order a cable, you get the wrong cable, you get different wrong crossover, and everyone every bender's pointing fingers at one another, and you trying to get the software up and, like, where am I right now?

Speaker 4:

And by the way by the way, just to set that I mean, the the audience generally at this event, if you pulled 10 of them and asked them just what, like, AWS is or what cloud computing is, would blank stare you.

Speaker 2:

I did. Totally. Yeah. So the the fact that this guy was, like, I had had pain with nicks.

Speaker 4:

In a data center.

Speaker 2:

In a data center was definitely, let alone. So he's definitely surprised. I'm like, well, I, so I start explaining Oxide, obviously. And he's like, terrible name. You have to change the name.

Speaker 2:

I'm like, okay. Well okay. Okay. We're not gonna we're not gonna do that. But, yeah, but that that one's off the table.

Speaker 2:

Like, let's, and I'm like, actually, let's go. Let me go show you the rack. And so I took him kinda back into the church and showed him the rack. And, and, you know, he it was really interesting because this guy was, caffeinated plus plus, was, very, had had technical experience doing this, but also really understood kinda like, you know, what the differentiators would be. And he he really felt like I had I had undersold it, apparently, to him.

Speaker 2:

He was, and and readily acknowledged that the branding was like, that was it's like, yeah. This looks sick. This is good. But he, in particular, he got to the back of the rack, and he's looking at the the cabled backplane, and he's like, are you fucking kidding me? Look at this thing.

Speaker 2:

Like, this is this is you know, it's he's really kind of, very exuberant, which is great. And it so I e, as he's kinda going through the things that were just mind bending to him. And so, like, in particular, he's talking about the cabling and the boxes. Like, how many boxes is this thing shipping? Oh, it ships in one large crate.

Speaker 2:

It's like, you gotta say that on your website. That's gonna be the topic of your crate. So then the the the blog entry was in no small part, inspired by kind of, caffeinated plus plus. This is actually, later in the conversation, he would later, reveal that he he may have been on things additional to caffeine, and, they have offered us to indulge Steve and I both to be heard. No.

Speaker 2:

Thank you. Strange. Yeah. Yeah. No.

Speaker 2:

Thank you on that one. But the, but thank you very much for I

Speaker 4:

just dropped a picture of it and see. Oh, did you?

Speaker 2:

Oh, nice. You did. Yeah. The the

Speaker 4:

long shot. The long shot

Speaker 2:

that is. I know it's just

Speaker 4:

it does look photoshopped.

Speaker 2:

It it does. It is not photoshopped. It.

Speaker 4:

Yeah. It looks extraordinarily photoshopped. That is an actual picture taken. And then, yeah, there there were some other some other photos that came across in that setting that were, you know, like prom worthy because of the back background.

Speaker 2:

And now Adam's thinking like, oh, great. Now you have to, like, make this the image of the podcast, or you gotta make me go in and do video editing again? That's right. I absolutely image.

Speaker 1:

I got it already. I'm also trimming this out, so we're fine.

Speaker 2:

Yeah. Yeah. There there we go. We're gonna trim out this whole discussion never happened. So that was kind of the the the motivator for getting into, you know, as we're because I think that realizing that we needed to speak about this really from that perspective of

Speaker 4:

and by the way, we it should be stated this was a this person is a pretty good marketer.

Speaker 2:

Yeah. For sure. Absolutely.

Speaker 4:

They have people spending 100 of 1,000,000 of dollars on canned water. Right, of which we may be drinking right now. Right?

Speaker 2:

It's good it's good canned water. Good canned water.

Speaker 4:

But, he had some good suggestions.

Speaker 2:

He had some really good suggestions. And I think that what and just in general, I think the just, capturing and what we tried to capture in the blog entry was the the what we have heard over and over again from folks, who when you see this, this is not what you are accustomed to. This does not look like a a a stack of commodity gear. This looks very different. And it looks very different in a bunch of different ways.

Speaker 2:

And so we kinda tried to capture a bunch of them, and talked about things like the the the cabling and the fans and the fact that it is not something that you're you're cobbling together in the software. But it's like, I also feel that that that, you know, try as we might, we can't capture the whole thing. So we're trying to capture some of those those first impressions. I think they told me to keep it to a 1,000 words. I I think I definitely did not do that, but it wasn't, you know, it's only over by scale.

Speaker 2:

Oh, that's just exactly close. But so that was the the kind of the genesis of the, of the blog entry, and the, yeah. I I actually think so there are people in the chat being like hold on. The you let me just get this straight. You, Brian Cantrell, are complaining about someone being excitable and seemingly and it's like, yes.

Speaker 2:

That is exact yes. You're just gonna have to, like, take me at my word here that this is I think and, Steve, you were talking to both of us at the same time. I think you can confirm that he was a faster talker. Yes. It was more amped.

Speaker 4:

Yes.

Speaker 2:

Okay. Thank you.

Speaker 4:

And and just absolutely zero off switch. So there's a Oh, what are

Speaker 2:

we talking about? Me or him right now?

Speaker 4:

He's he's just a every night. Didn't really have any Oscars. Right? No. He, I mean, he did not get there was not a room to get a word in edgewise.

Speaker 4:

And there were folks there that definitely were, you know, kind of trying to break in

Speaker 2:

and No. You'd it was no. Yeah. It was, so, yeah, maybe this is what it's like talking to me, but it it's like, hey. Look.

Speaker 2:

It's exhausting. Okay? I've got I I you all have my my sympathy. It's it is, it was a lot, but he was also right about a lot of things. So, you know, so it's fine to get it out there.

Speaker 2:

So we had the blog entry written, and, so that's gonna we're gonna turn the key on that one at 3 AM Pacific.

Speaker 4:

I like We we have folks in the UK that easily could turn said key.

Speaker 2:

Yeah. Yeah. Yeah. Yeah.

Speaker 4:

It would be later in the day. Yes. We can instead, I think you were up at

Speaker 2:

I was I 3 AM. It it that sounds like a criticism. No. Your president. No.

Speaker 1:

It's just an observation.

Speaker 2:

Right. I I was like, look. If I wake up naturally, I wake up. And I just happen to wake up at 25. I don't know.

Speaker 2:

My, you know, my body knew. My body knew it was home. I was I, Yeah. I don't know. I got some, like, yeah, some hacker news.

Speaker 2:

Like, I don't know, some hormones or something. The, so got that thing kicked at, at 3. Got the blog entry up. And I I mean, so, Steve, and so, Clark, do do you because I do not submit any of our stories. I I always want that to happen kind of organically.

Speaker 2:

Is that your I mean, as as someone who's a top 25 in top

Speaker 3:

post everything that I believe should be posted as soon as Oh, do you really? Oh, shit. Yeah. Well, it's not not as soon as you see it. I think that the correct time to post in general is at, 9 AM PST, and that is because you get the people who are just showing up to the office who don't wanna start doing any work and they're drinking coffee and they're reading Hacker News, And you get the people on the East Coast who are at lunch, who do not wanna do any work, and are looking at Hacker News.

Speaker 3:

And I think that helps you get the most chance of being uploaded to the front page.

Speaker 2:

Of being uploaded to the front. And and so just for those of you who do not spend as much time your lives online, you're living much better lives than than Steve and I are. So with, I mean, with Hacker News, it's like you don't really control it. Right? I mean, people are it's gonna get outvoted, but if it's kinda compelling.

Speaker 2:

If it is, I I have had this happen where I submit things that I think are super interesting that get zero attention, and then they if they kinda don't survive in that kind of that new bin, they, they will just never see any more close to the front page. And on some of those, they will actually, the the folks at Hacker News will say, like, hey. This is really technically interesting. This is important to the audience. So we will later give this an artificial boost, which is great.

Speaker 2:

So that's always an option on on on stuff like this. But yeah. Steve, generally, I'm with you. Like, 9 AM Pacific is a good time to do things. I would say 3 AM Pacific is not the time I would pick.

Speaker 2:

So

Speaker 3:

Yeah. I actually so I had asked on our pull request on the draft of the blog post. So I was like, oh, by the way, the way is this going out? And Ben said, oh, 3 AM Pacific time on 26th. And I was like, oh, cool.

Speaker 3:

Because, like, Ben's a funny guy, and, he makes jokes sometimes. And I thought that he was, like, joking about it going out at 3 AM.

Speaker 2:

Good one.

Speaker 3:

Like, oh, that's really funny. And I just didn't actually follow-up. And then, like, I wake up and I was, like, man. I'm still pretty sleepy. And then I'm, like, let me check with the computer.

Speaker 3:

And I was like, wait. Holy crap. It's already there. There's a bunch of and then I was like, oh, Ben wasn't joking actually. It's been up for 3 hour.

Speaker 3:

And, like, immediately got involved. But, yes.

Speaker 2:

That it was not sarcasm. Yeah.

Speaker 3:

Very funny. Yeah.

Speaker 2:

Yeah. So the so, again, I didn't submit it, and, but, you know, as we're kinda getting everything else out there because, like, you gotta get, like, all the social posts out there. So I'm posting to Twitter, and I'm realizing that, like, you know what? I don't have the blue sky creds for the Oxide account. Yeah.

Speaker 2:

And I kinda don't care. I think this is, like, oh, blue sky. This is, like, you you know, you know, like, your true social network is the one you check after an earthquake. And the fact that, like, I couldn't even be bothered to I'm, like, sorry, blue sky. I don't know.

Speaker 2:

I I'll I'll deal with you later. Blue sky's gonna open up completely to the public. It's the problem with sky. It's still, anyway, and I'm, I'm still optimistic about blue sky, but I did get the the post up certainly on on Twitter, and on Mastodon. I mean, I think I think it's super interesting.

Speaker 2:

And, Steve, I don't know if you found this for your stuff as well. For tech audiences broadly have moved to Mastodon, and the engagement on Mastodon is better. It's like you almost want I mean, we, so really important to get up on Mastodon. And then it is really important to get it up on what is rapidly becoming the 3rd most important stuff in that work. I cannot believe say this.

Speaker 2:

LinkedIn. LinkedIn is like I feel like the big story about the the the fact that Musk has flown Twitter into the side of the mountain is that LinkedIn has become actually way more relevant. The content is better. The engagement is good. They don't have, like I mean and now that said, like, LinkedIn has got all the same problems that LinkedIn has always had.

Speaker 2:

Yes. And I'm wondering if LinkedIn has

Speaker 3:

What is it all

Speaker 4:

things over

Speaker 3:

there? LinkedIn. Everybody that came up to our booth was, like, hey. Can I connect with you on LinkedIn? And, like, the first time it happened, I, like, laughed, and I was like, oh, yeah.

Speaker 3:

Sorry. I don't, like, have that. And the second time it happened, I was like, what the fuck is up with LinkedIn? And then the third time, I was like, am I gonna have to resign up for LinkedIn? Like, why is everybody so interested in LinkedIn now?

Speaker 3:

But, yeah, maybe you're very old.

Speaker 2:

Well, and I think it is because it is beginning to serve, you know, Twitter was serving this kind of professional capacity for a lot of folks. And I think it is, LinkedIn is in Mastodon kind of between the 2 of them. Anyway, so I I've, like, I didn't forget this LinkedIn post up there for kinda get it. Make sure to post it. And we actually, actually, a lot of great engagement on

Speaker 4:

on LinkedIn, Facebook for professionals. That is a good it's a good descriptor.

Speaker 2:

It oh, it it it for sure is. And it's it With

Speaker 4:

all the trash All the trash. Somebody to benefit. Right. That's right. It's like, don't call Facebook.

Speaker 4:

Oh, good. No. Just that it, like, it it sadly is still an easy way for your extended friends and family to keep up with things that are happening when you got kids. Like, it's easy to have a vector where it doesn't all have to be telephone calls and texts. And Facebook's terrible, but it serves that purpose.

Speaker 4:

It does. LinkedIn, similarly, is this place where you've got, you know, your your connections from your professional career, and it is a place where you can update people.

Speaker 2:

Well and I think and and then the quality of engagement has gone up as Twitter as because you have, and I know Steve right now, you can't see his eyes, but he's just like, I I I hear your words, but I can't actually process that. It actually is the the engagement is is better because I think it, anyway, it's replaced Twitter. It's they will get it up on LinkedIn, bunch of engagement. And by the time I get all those up there and I, like, spend, I don't know, 10 seconds trying to find the blue sky credentials, and I'm not able to do it, I was like, oh my god. This is already on Hacker News, and it's the number 2 story.

Speaker 2:

And I'm like, wow. Okay. I guess we're coming in hot there. And then I, like, click reload, and it's the number one story. I'm like, alright.

Speaker 2:

Well, here it is. It's like, at this point, you know, 3:30 in the morning Pacific. And I wanna go back to bed. I do not wanna sit here, and, but for those of you I'm sure, you know, anytime you got, you know, the Internet talking about something you've been a part of, it is hard to not, you know, click reload on it. It's hard not to stare at it.

Speaker 2:

And

Speaker 3:

It's also big in hacker news, like, for people who don't spend their lives posting a hacker news. First of all, congratulations. And second of all, like, the initial comments can really set the tone for, like, everything else. And so it's, like, semi critical to, like, shape the conversation in the way that you want it to go early on in the thread's history because, like, otherwise, it could go to, like, literally anything. And,

Speaker 2:

like Yeah. I think you're totally right, Steve. And I think that you and I have both seen this where if somebody sets the tone early, then it's gonna attract a lot of negativity, and people just aren't gonna engage with it because it's gonna be I mean, because often if if there is if people actually wanna engage with something positive, positive about it, or they got questions, so they wanna get kind of technical answers, and they can just see that's like, alright. This is just a mess in the comments. They'll just walk away from it.

Speaker 2:

So it's really important to get kind of an initial comment out there that people can kinda latch onto it if, and upvote if if they so if they so feel it. So you really wanna as as Janet was saying in the chat, you kinda kinda set those thread vibes early. But look at the thread vibes are good. I have the thread. I'm like, god.

Speaker 2:

This is all good. This is all basically tacky. This is not nothing is too far off the rails here. And, you know, whenever you've got someone saying kinda something nutty, you've got someone else coming in and answering them. So it's like, alright, this is good.

Speaker 2:

And I'm like, alright, I'm gonna I'm gonna go to back to bed, actually. I'm gonna I which is not something I would have thought that I was biologically capable of doing. It's like yeah. It's like, I I am actually going to sleep with us as the number one story in Akron News.

Speaker 4:

Being up at 3 AM? Yeah. I at least surprising going back to bed.

Speaker 1:

I know. I'm good. And it's and just curious. That I don't know. Like, did it really happen that way?

Speaker 2:

Maybe. Right. Exactly. But maybe not. Yeah.

Speaker 2:

As far as you know.

Speaker 3:

That's right.

Speaker 2:

Yes. Look. I may have just been clicking reload and keeping a low profile. He's calling after 1, and I went to sleep just to but it was good. So the the comments are good, and so I think we wanna, like, get in, to, the I get this right.

Speaker 2:

Someone's saying the chat, like, okay. He actually is aging. I assure you I am aging, and if you would like, I can do a my my children would be happy to to point out all the elements in which I'm aging. My 19 year old can't run me, which is very I don't know. I don't know if it's with that.

Speaker 1:

That's tough. No. Better not be lying around

Speaker 2:

with them

Speaker 3:

in a bear anytime soon or you're gonna be in trouble.

Speaker 2:

That's exactly right. I know. I I need to make sure that I I I take my younger children with me who I can still outrun, fortunately. But so we there was a bunch of questions in there. I think that we and then a bunch of new stories.

Speaker 2:

So the, I I think we wanna kinda the news stories, I think, were less surprising. I mean, I think that they were and especially folks who have been following Oxide, I don't think you're gonna find any big reveals. One thing that I I did wanna, but I thought maybe we could go through some of the the big themes on the questions and things that we did not. So when we when we want when we we ship first shipped, in July, end of June, July, we did a podcast where we kinda answered questions. And so I would definitely refer people to that.

Speaker 2:

I mean, because we answered a bunch of questions there. I don't wanna go through and kinda because there were some of the certainly some of the same questions. I'll but there are also some new ones and some important ones. One of the things that I was surprised by, and I I would love to get everyone else's take on it, I was definitely surprised by the number of people who are like, this is not cloud. Cloud has to be rented.

Speaker 2:

Like, that's part of the definition of cloud. I go, okay. And somewhat and I I think I did see this comment before I went to bed. Maybe this is where I was able to send me to bed with with some confidence, is someone had quoted the NIST definition for cloud. And this is your like, oh, man.

Speaker 2:

I hope NIST has got the right definition here. I I hope I hope you all spread a little bit. Like, oh, this maybe we're maybe we don't have this link to go to that. Exactly. And the NIST definition actually, as with many things NIST is really good, actually.

Speaker 2:

The the NIST definition of cloud computing is like right on. Spot on. Spot on. I mean, a really good definition.

Speaker 4:

You have it?

Speaker 2:

Yeah. It does. Adam, drop it in the channel. And and it it is just, you know yeah. The cloud computing is a model for enabling convenient on demand network access to a shared pool of configurable computing resources that can be rapidly provisioned and released with minimal management effort or service provider interaction.

Speaker 2:

That is a so Well done, NIST. Well done, NIST. And whatever technologist, I mean obviously there are technologists either at NIST or friends of NIST that help write that definition, and I know you may feel that your job is thankless. We'd like to thank you here at Hawkside because I, Adam, dude, that was, I thought that was a really good definition. What'd you think about

Speaker 1:

I think it's really good. I think, you know, I'm I'm tempted to start a talk with, you know, NIST defines cloud computing to me.

Speaker 3:

I was gonna joke about a best man speech that starts on a wedding reception.

Speaker 2:

This defines love as Yeah. Right? The but I think that was really good. So with the, and, so that was great. And then so that comment, that kind of, like, angle of, like, no.

Speaker 2:

No. This is not cloud cloud. It has to be rented. That just got absolutely, like, launched into oblivion. It was really great.

Speaker 2:

And then there's, like, one of the comments in reply to this is, like, yeah. Sorry. You're just gonna get down voted here because, like, you're just wrong. Like, nobody agrees with your definition, basically. So it was good.

Speaker 2:

Maybe that's I think that was the lullaby that I needed. There you go. That Off to bed. Off to bed for me. This is great.

Speaker 1:

So so, the reception was good. Do do you wanna get into some of I pulled out some of the comments. Do you wanna get into some of those from from Hacker News?

Speaker 2:

Well, so one of the I do wanna, like, just tack directly into one of the themes that I because I I I think that something we saw a decent amount was like, hey. How much does it cost? Like, the fact that, like, their pricing is contact sales is like, oh, okay. Yeah. I've seen this before.

Speaker 4:

When that came up with reporters as well. Yes. Because, obviously, reporters want to get as much tangible information that they can go right on. And so we had that question also posed as we were doing the interview process.

Speaker 2:

Yes. And one reporter in particular was, like, asking us about pricing, and you gave them the answer that's the honest answer, which is It depends. It depends. People do not wanna hear the it depends answer. Like, oh, it depends.

Speaker 2:

You're being

Speaker 4:

I see.

Speaker 2:

I see. You're being evasive. I got you.

Speaker 1:

Means it depends on how much money you have. Like, how big is your budget?

Speaker 2:

That's how much the comp price is. Yeah. Exactly. What's your what's your timeline?

Speaker 4:

What's your time Budget. Right. Yeah. No. It depends because while this is a rack scale computer, there are things that are variable.

Speaker 4:

So one is the number of sleds that you have in this rack scale computer. You can buy it complete with 32 sleds. You can initially install it with fewer than 32 sleds. And we've had customers that have needed to do that for reasons like we wanna get this in under a certain power footprint, and then we wanna expand it as we see how much we're using. Or we have a budget limitation that is going to expire, and so we wanna get it in and then be able to grow it from there.

Speaker 4:

And so in that case, when someone is getting it in for, you know, fewer sleds than 32, the the system itself can come in below $1,000,000, which is what I what I told him.

Speaker 2:

Right. It is what you told him.

Speaker 4:

Print it. He didn't print he wasn't very interested

Speaker 2:

in that, I guess.

Speaker 4:

He's like, no. I Because and then I and I said, and and it can be above $1,000,000 if you if you're configured a certain way. Because you also have a decision you can make about how much DRAM, how much memory you want per sled. We've got 16 dim channels. You can have half populated to fully populated, so you can have half terabyte or terabyte of DRAM per sled.

Speaker 4:

And so you've got these kinda 2 dials, and he, yeah, he was definitely not interested in the nuance. He's like, what is the price? Right. Right.

Speaker 2:

And so that was I and I I I'm definitely getting from the chat. We will never again say contact sales. Obviously, contact sales is clearly the all the antibodies come out. Look. We will need to find some some synonym for reaching out to sales so we can give you I don't know.

Speaker 2:

The

Speaker 4:

Yeah. That that it's tough. It's very fair.

Speaker 2:

It's a fair question. So the, I I think that, and I think people, you know, you wanna kinda know what the price is, but it it it which is important. And, you know, we obviously need to have a little a little configurator on there so you can kinda it and you can then see with it the and get an actual quote, which I think is important. I

Speaker 4:

think that the, Well, especially because what folks wanna do is they want to be able to kind of put together a comparison.

Speaker 2:

They wanna put together a comparison.

Speaker 4:

Totally reasonable. You know, right now, you can go to Dell's website, and you can configure a server that has the same components. So you would put in, like, 103.4 terabyte NVMe drives and 16 dim slot like, 16 dims of 64 gig memory.

Speaker 2:

I like the way you are instructing the future listener to be like, did you get that 10? No. No. No. Not 5.

Speaker 2:

You need to put in 10 if you wanna get the accurate price. 16, get the numbers right.

Speaker 4:

They find that can't go higher than 8, but you, and then, you know, an AMD processor. And if you're doing that, and then if you're adding things like, for example, VMware, it will accurately flag you that you need not one VMware license. You need 2 if you selected more than 32 CPUs. But, anyway, it'll error out, and then you can kinda correct that, and it will give you your your your final total. And I think, it it's it's a pretty eye opening experience because it just helps folks get a good sense of what these things cost.

Speaker 2:

And we I think it's fair to say we had a reporter who made it who remained nameless, who made a bit of a mistake, and didn't do exactly what you prescribed, And instead, got a a quote for 32 CPUs.

Speaker 4:

When you say CPUs, you mean Like a server with 32 CPUs? I'm glad you asked.

Speaker 2:

Okay. No. I mean 32 CPUs, unsocketed CPUs.

Speaker 4:

Like a bag of chips.

Speaker 2:

Yeah. Yes. A bag of chips. Right. I I mean, 32 c the it's like actually be very you might be surprised to learn that that, what you just ordered can actually go to a backpack.

Speaker 2:

It's actually it's a heavy backpack, but it's basically a backpack. And the, the the number that you get is, of course, a lower number because, you have actually you've you've priced out a car against the cost of the tires. So, we we did correct that. We could we we, we encouraged the reporter to do and maybe walk the reporter through the way to actually get a process started.

Speaker 4:

There are fewer and fewer people today that want to then, make their own PCB and do their own assembly. And Right. You know? Right. Buy buy buy a bag of network cards and DRAM.

Speaker 4:

And so, yeah, it it it, but, again We

Speaker 2:

we got it corrected, though.

Speaker 4:

Got it got it corrected. Got to a good spot. But it's important.

Speaker 2:

And in fact, it got that obviously in configured on,

Speaker 4:

properly.

Speaker 2:

It got got a kind of a proper price comparison, and I think that, anyone that does a proper price comparison is gonna discover, like, wow. Okay. Actually, this is actually very price competitive. This is not, so we we we do not wanna be you know, we're a very transparent company, and I I I think that we can certainly understand how people felt we were being evasive by contact sales, but that is not what we're trying to do at

Speaker 4:

all. No. It used to be I think that the the original link was sign up for a a spot in line, and maybe we should revert to that.

Speaker 2:

Yeah. Yeah. Yeah. Yeah. Yeah.

Speaker 2:

Or or just the I know Or contact us or yeah. I don't there's there's plenty

Speaker 4:

of better ways to to, phrase that.

Speaker 2:

We we gotta find a way to phrase that. Anyway, the, we again, we're not trying to be evasive with the pricing. So we'll, avoid

Speaker 4:

Not at all. Because, again, it's it there's a bunch of public information about, like, what does it cost today to put the equivalent amount of CPU DRAM storage into a rack, a commercial rack out there, and then, warranty it for a couple of years, and then get soft pay for software to to be able to to manage it and organize it. And, as as we had customers tell us early on, folks that that were running data center infrastructure, it's like, yeah, a rack full year with all the software licensing over 3 years cost us you know, I think the the the last one we heard from someone was, like, $2,100,000. Right? And so we yeah.

Speaker 4:

We we're we're gonna get the the numbers out there more broadly. And with this particular reporter we had just shared, like, can be just under a1000000, can be over a1000000, and depends on the configuration.

Speaker 2:

Yeah. And and I think what you you can I the question's in the chat of, like, hey? How hard would it be for you all to do a configurator? And the answer is not hard. That's something we need to go do.

Speaker 1:

So No.

Speaker 4:

And we I mean, we've we'd we did a bunch of the the the groundwork on that, because as you can imagine, when you're going through things like a fundraise, investors wanna understand. Well, what is what does this look like compared to someone running a bunch of infrastructure on AWS EC 2?

Speaker 2:

Yeah. Yeah. Yeah.

Speaker 4:

Right. And so we went through, and, you know, there's a lot of things you need to do to do a fair and accurate comparison. You you start with pick an arbitrary e c two instance or a couple of them and then figure out how you could pack that into some subset of the capacity in the rack. Then you have to fairly do some balance of on demand and reserve pricing. Right.

Speaker 2:

Which is a total I mean, that is a bit of a trick. Right? Because there's a big, big difference between on demand and reserve pricing. And so, you know, this is where, you know, all kudos to AWS and their ability to execute. But the fact that that reserve pricing is so much cheaper shows you that there's real hardware behind it.

Speaker 2:

It's like Amazon needs to buy stuff to run this stuff on, and they need and they're gonna give you if you want a true, completely elastic, I'm gonna give you zero forward visibility to how much compute I'm gonna use, and I wanna use as much as I as I possibly can. One, you're gonna find that there are actual real limits, and you go to permission. You're not gonna be bridging 10,000 VMs. You're gonna be bridging quite a bit less than that before you get kind of a, you need to call us. But you will also be absolutely paying for that, and you're you'll be paying a lot for it.

Speaker 2:

So that, and thank you very much folks in the chat for helping us brainstorm on ways to be transparent about our pricing. So and and, I like folks, like, don't forget, like, the cost or the power to to indicate the which is true. So the another thing that kinda came up, and, Adam, I don't know if you saw that that something I definitely went back and forth on a little bit on the acronyms comments of, in a in a very good way, but about, about the fans, and about the noise, and why the noise is actually representative of of a much of much bigger things. That it's actually, like, we weren't actually trying to design this to be, for its acoustics. That was an artifact of of the way we'd built this thing, and and took a clean sheet of paper.

Speaker 1:

Yeah. The fans are a are representative of what I think are, some a part of the value proposition that we don't talk enough about. But we do talk about the fans. I know you love your 80,000,000 millimeter fans, but the impact to power consumption is pretty wild. I mean, really significant.

Speaker 2:

It is. And the and right now and I think the the the the fans okay. Look. I know people in the chat are being like, like, oh my oh my god.

Speaker 1:

The fans the fans. Right.

Speaker 2:

Oh god. Will the senator from 80 millimeter fans please yield? So, you know, the fans we have not been talking enough about are the fans on the AC power supplies on the one use and two use. I feel like those fans, I have now We we

Speaker 4:

we are back in 10 minutes if you wanna skip the hours of

Speaker 2:

okay. We're done. So wow. Okay. Anyway, that's that.

Speaker 2:

That's the end of that. The these because you have these little teensy tiny fans in the AC power supplies, and those things are also screaming. And this is where you get to the and actually, we did not factor that into our calculus for I mean, we just kinda knew we were gonna use the DC bus bar, but we did the the calculus anyway. But I think that the the power consumption for the fans in the power supply is something you've got zero visibility into. I don't know.

Speaker 2:

Every single power supply that like an AC power supply that you're gonna put in a commodity server that's gonna give you that visibility. Alright, everybody. That was a fast 10 minutes. That was a fast 10 minutes.

Speaker 1:

But it's so so how significant percent percent comparison. It's good.

Speaker 2:

Yeah. So so how

Speaker 1:

how big I mean, how how much is the do these fans draw? Like, what's the power difference?

Speaker 2:

It's on the order of 20 to 30% of the rack is the I mean and so we are and I so we say 2%. It's actually not 2%. I say under

Speaker 1:

2% of our

Speaker 2:

the 2% is super conservative. It is, like, pretty far under 2%.

Speaker 4:

And here we go again. But it is but but but, I mean so these 1 u fans can take up the the combined fans 20% of the power of the times n number of racks. This ends up being a huge, huge, huge inefficiency. And getting that from 20 plus percent to under 2% just means way more density.

Speaker 2:

That's right. I think it so this so, Steve, this is an important point you're making because it's like people like, woah. I don't care about my power bill that much. So that may be true, but you actually do care about the you actually do care about the amount of compute you can put in per unit, per per square foot, per tile, in your DC, and that's what this goes to. It goes to density.

Speaker 2:

This all goes back into compute, and you want it to be used for compute and networking. Another, another question that came up, and, oh, I should say, we are gonna be doing a kind of a series because someone was asking about, you guys need to do a a blog post on the the fan power. I can't tell if that's, like, a so I don't have to read it as opposed to, like, listen to it. I can't this is kinda like when I tell, like, I tell my mom that she should write a letter to the municipal authority that she's so upset about. It's like, so I'm not I'd be I'm not sure if I'm gonna pull the right letter.

Speaker 3:

Like an an oxide only fans that just post pictures of the fans would do pretty well. I'm just saying. Still.

Speaker 2:

Oh, man.

Speaker 3:

While we're talking about social networks, quote unquote.

Speaker 2:

Wow. Pandora's box has been opened. I that is so dangerous. I the you know, yeah. Exactly.

Speaker 2:

Some ideas I should not have heard. That that's Steve, I have you have you

Speaker 4:

you've been sitting on that for help? Yeah.

Speaker 2:

You've been sitting on that one.

Speaker 3:

I feel like that joke has been made several times,

Speaker 4:

times, but Oh, I don't think so.

Speaker 3:

Maybe not in this call anyway, but, like, it's fine. It's fine. Anyway.

Speaker 2:

So, yeah, that well, okay. But I'm just kinda blunt, but, yeah, I need I need to recover from the only fans. The but we're gonna have a, a series of blog posts, and we actually wanna do episodes of Oxide and Friends where we go into some of the stuff in detail. We'll we'll save the fans for just the blog post, but we wanna talk about a a bunch of different, elements of the rack, and go into detail there. So so where and we obviously got an idea of the detail that we wanna hit, but if people wanna see particular details, we'd love to hear from you.

Speaker 2:

Make sure to get an episode teed up, and make sure to to get a discussion on it. So, in in a blog post. So, Adam, what were some of the other big themes that you Yeah.

Speaker 1:

Some of the some of the other themes, tacking into value proposition. I think that there were a bunch of commenters of the form. Hey. Can't I just get all this stuff from Supermicro and load it up with OpenStack and do it all myself? And then a bunch of other folks saying, let me know how that's working out for you.

Speaker 1:

And, how do you get all this thing working together? And you've got 6 different vendors in the mix. And when there's a problem, how does that work out?

Speaker 4:

And even further than let me know how that works out for you. But you had folks that were going in and saying, let me tell you my story. And then telling their own stories about unboxing and assembling and integrating and troubleshooting. And back to the comment of this this person that was at the, VC event. It's like, you've done all this, and then you find that you have the wrong rail.

Speaker 4:

One wrong set of rails Right. Sets you back another 5 weeks. Right? And even once you get to that point of integrating all the 5 vendors together, then, Adam, to your point, folks commenting in there, like, well then, what are you gonna do when you have a problem? Who who are you calling?

Speaker 4:

And which which which of the vendors are you finally pinning down that are pointing at the other one? And I think it was it was great to for us not having to chime in and others chiming in.

Speaker 2:

Others chiming in. And experiences. I I did like there was there was someone, in the Hacker News comments who, speaking about Dell in particular, said that and then Ryan is quoting this, that Dell is a single vendor that will diagnose and fix all of your hardware issues. Just like full sentence, no further qualifiers, which I I mean, this is a a troll in the in the oldest sense.

Speaker 1:

You could practically hear the guffaws and the replies to that.

Speaker 2:

Well, I love the fact that Steve Klapnick, you said that you were with a a a friend of yours or an acquaintance of yours who has deployed, like, large amounts of Dell at, like, a social event, and you're, like, reading him hacker news comments just to get him, like, guffawing.

Speaker 3:

Yeah. Yeah. I I like, one of my one of my neighbors, works for the city, and they have a pretty large deployment of, like, VxRail apparently, in whatever his department is. And, yeah, I was just like, yeah, dealing with that support is absolutely terrible.

Speaker 2:

Right. It is, it's not great. And so and you get these fingers pointed at one another, And I think that, you know, with OpenStack, it's, OpenStack's not a product, which is part of its problem. And, you know, back in the day, as kinda more and more companies were throwing their lot in with OpenStack, one of the questions that you get is, like, well, how are you gonna go compete with all these hundreds of companies now that are working on OpenStack? And it's, like, are you, like, can you hear yourself?

Speaker 2:

I mean, the last time that that hundreds of companies cooperated on something and had a successful result is like World War 2. I just don't I mean, like, you you can't have one company do something without getting in its own way. Let alone like, oh, no, don't worry. We have many different organizations that are it's like, okay, yeah.

Speaker 1:

Yeah, I really appreciated how much folks got the value proposition of one entity responsible for this whole thing from networking and CPU and virtualization and all this stuff. Because I also think that that's a huge part of why we built it the way we built it. You know, why we are running the operating system or running, why we're building our own service processor, our own embedded OS. It all it all comes from that desire to support this whole thing, this whole, you know, box full of lots of heterogeneous components that we are integrating.

Speaker 2:

That's right, Adam. And I think that, you know, in every one of those decisions was happening, not because we were, like, electing to do something on our own, or go our own way, but because we had had specific concrete experience where the inability to kinda control both sides of this led to a suboptimal, to put it mildly, experience.

Speaker 4:

Well and I think it it the thing that was great to hear come through in some of those responses and acronyms was, you know, a vendor that is saying, you know, hey. This vendor will support, fill in the blank, their product. And folks being like, yeah. Well, you know, if I could have diagnosed that it lived in their product already, wouldn't that have been nice? Yeah.

Speaker 4:

And instead being in a position where, you know, someone was saying, like, you know, I whatever issue I have sitting above this entire stack, I can throw an oxide. And that's true. That's true.

Speaker 3:

It it could be

Speaker 4:

an issue in your database. It could be an issue Yes. Wherever it starts from, we're gonna go run that to ground. Adam, to your point, be because we can, because we control the whole stack. And because we've been on the other side of the table in that helpless position where people are kinda bouncing you all over the place, we will.

Speaker 4:

Like, we're gonna take ownership of every one of those issues. And, it's it's not as sexy as a speed or a feed, but it's folks that have actually run infrastructure on prem. Hearing them come through and telling those stories, it means as much as any product feature you can have.

Speaker 2:

And it's really, really important to us. I think it it does kinda, hit on 2 other themes that kinda came up in the Hacker News comments. One of which, and this article also came up in the chat. So I was like, did you all forget to make the Helios repo public? And the answer to that is basically yes.

Speaker 1:

And what is Helios? About that.

Speaker 2:

What's that?

Speaker 1:

And do you wanna describe what Helios is?

Speaker 2:

Yeah. Helios is so it's our Alumos derivative. The Alumos derivative itself is is actually all public. That's in the Saint Louis branch of our Lumos fork. So that is actually the the meat there is all actually public.

Speaker 2:

But and as I kind of was replying to that on hackers coming, someone was like, how was I supposed to find that? I'm like, yeah. That's not on you. You were not supposed to find that. That was like, it's, so sorry about that.

Speaker 2:

But the that we'll be rectifying that, which needs to be there's a small amount of cleanup that needs to be done. Like, we do need to add a license in there, before we but the actual Helios repo really just contains the things that kinda point to these other repos. It's not actually it, it it kinda pulls it together and builds it. But the it is very important for us, and, you know, a big differentiator for us is the fact that it's all open. So from from top to bottom, we really, have wanted from the beginning to get it all out there.

Speaker 2:

And, you know, one of the things that actually, you know, something that people didn't ask us about, that I think maybe a question we kinda what's that, Steve?

Speaker 3:

Oh, nope. Sorry. You're in the middle of the thing. I'm a try to interrupt you. Saint Louis is actually also not public.

Speaker 3:

So saying that it's on a a branch on Saint Louis means that you could find it. Like, that's also not open,

Speaker 2:

Oh, and that okay. Well okay. Yeah. We I'd so we'll we'll reconcile that. Yeah.

Speaker 2:

Exactly. But, you know, the, one of the questions that that we would ask our ask of ourselves is, like, hey. What have you learned from your first couple customers? Like, what has been the the experience been, and how have things, you know, you've got some kind of some some theories as you are getting this product in the market, and how did those theories work out? And I would say that we've we've got a lot of proof points of, like, yeah, we've got, you know, I we were having, one tricky problem with a customer just with the, the necessary kind of cert generation that was not actually like it it it was at I mean, if you wanna get super technical about it, it wasn't necessarily an oxide problem, but, like, was definitely a problem, like, getting, like, we needed to pull this together.

Speaker 2:

Yeah. To get it up and running. And, you know, the the the line that we heard was, like, you know, with with VMware, I feel like I would be kinda, like, on my own right now. Even though it's like, no. No.

Speaker 2:

We're not gonna be it's it's really important.

Speaker 4:

Not just on my own. It was, like, while you all were troubleshooting this, which to be clear, this is because we, we integrate into someone's identity and auth provider. So we don't try to replace that. We we leverage, like, the SAML provider they, or or the fee that they use for identity and and auth. And so they have to give us certificates.

Speaker 4:

And in the in the process of getting that up, we had them on our troubleshooting call. They're like, it's so refreshing that I was allowed to just sit back and kind of listen to how you're working through this problem. With VMware, I would have been, like, handed an SDK, and they would have disappeared. Or I would have been, you know, sent to a, you know, online support system, and it just you know, weeks later, maybe I emerged from something. And it was it was good to hear.

Speaker 4:

And and having the ability for folks to for that that is open source, which we are quickly resolving to get all of it out there. Another comment was, they we were talking with this customer about they had some priority work that they wanted us focused on, and then there was in another issue that was low priority. And we just were updating them. Like, hey. By the way, you know, we're still working on this other thing too, and they're like, oh, I know.

Speaker 4:

I'm watching the bugs.

Speaker 2:

I'm I'm watching the issue right now. Yeah.

Speaker 4:

Exactly. I saw you bump that yesterday. And and then said, like, it's such it's so nice that I can kind of self-service in that manner and see what you what you guys are doing, because everything's open source.

Speaker 2:

So that's been, I think, really validating. And I think we've also had folks that have, been like, oh, yeah. Actually, I I pulled over a clone of this thing, and I I found, you know, an issue in this or I held so that's been, it's been, I think, really vindicating and and and validating

Speaker 1:

on that. One of my favorites was our customer was trying to run our CLI. We've downloaded a binary on some different version of Linux than the CLI had been built on. And so they just cloned the repo and rebuilt it and they had it, which is not obviously the best user experience, but the fact that everything is out there means they are empowered to do those things, which is great.

Speaker 2:

Yeah. And I think all of this is to, like, is it it's our empathy with folks that are running this thing. It's like we, you know, we we wanna treat you the way we wanna be treated. So that's been so far, and then we've we've we've made a bunch of specific technical improvements too that have been really exciting and to really improve that install time and really getting to the the point where I mean, I think we're awfully close to this vision of being able to get this rack rack in, powered on, and provisioning VMs within hours. You know?

Speaker 2:

And we're not that far off.

Speaker 4:

You were disappointed, and we would share the disappointment with the customers like, it's, you know, two and a half days. Right. They're like, this is exponentially faster than any other installation and stand up we would do. And then, of course, we're like, yeah. I know.

Speaker 4:

We're trying to get it to one day.

Speaker 2:

And we want to get to hours. A couple hours. Exactly. And the and part of the blocker there is, like, there is obviously a a super important and complicated, frankly, interface in with the customer's network, and so there's a lot that can go wrong there. So, the, I one of the questions that was asked, just in the in the chat here, if you had to build and ship the product again from scratch, what would you do differently?

Speaker 2:

Good question. It's a really good question. And it's really hard to I mean, I I don't know. I feel that we book our chamber time

Speaker 4:

a little earlier for compliance.

Speaker 2:

That's true. Yeah.

Speaker 1:

Higher RPEs a little earlier? I don't know. I mean, I don't think there's any art anything glaring architecturally that we regret or that we could have done differently based on what we knew?

Speaker 2:

There is some super small stuff, but a lot of that stuff too, we actually got right on on later iterations. And it would have been kinda, I guess, nicer to be able to but I don't I don't think it's really fair when you're going back and kinda thinking about that. I don't think you're allowed to, like, think omniscience. If you like like, I we we would've I mean, I mean, like, probably not on

Speaker 4:

we went on the path of using third party stuff. Yeah. Because we did a lot of investigation. And, I mean, if we again, if we get beyond impression, we can go back and and know that this is the wrong path, that we have to do this thing ourselves.

Speaker 2:

You know, you gotta go down those paths. I mean, I remember the the you know? And that was a really tough time when you, would you know, I remember, like, 2020, and it's the pandemic and everything else, and we're having, like, these kinda endless calls with perspective partners. And then, like, you would have a call that was a was like, we definitely don't wanna do work with this person ever, this company. And I remember Ari and a kind of because we would do these calls.

Speaker 2:

You remember those calls we would do these, like, debrief calls that you talk about therapy, where you'd be on one of these calls. It's just like, oh my god. That was like I I we had, like okay. GoToMeeting is terrible. This is I and I would like to say we had someone who had misconfigured GoToMeeting to be even worse.

Speaker 1:

I love your hot takes. LinkedIn, great. GoToMeeting, terrible.

Speaker 2:

That's right. That's right. Yes. Exactly. I feel like these are these are my hot takes circa 2,005.

Speaker 2:

That's true. So the but this is I I don't know if you were on this meeting, but you would go to the only way to unmute yourself was to use the keypad. And so you have to, like, star 8 to unmute yourself, and then it would be like, you are unmuted. It's like, well, I'm also like, okay. But, like, the person I was gonna interrupt or whatever, that's like, no.

Speaker 2:

The time has now passed. So, it was, but I I remember that, like, you get off these calls, you're like, oh my god. What a waste of time it was. And Arian would I I just remember him vividly saying, you know what? That was important.

Speaker 2:

We had to do that. Because we had to like, you could not have said that, like, the these folks actually do make products that are kind of viable for us, and we didn't know they were turkeys until we got them on the phone. So it's like, we actually had to do that. We can cross them off the list. We never need to talk to them again, but we had to spend that 3 hours or 2 hours, whatever it was.

Speaker 2:

And

Speaker 4:

the other one I could say is, like, you know, knowing that we absolutely could not build this off a reference design from a PCB version.

Speaker 2:

The other thing I would like Yeah.

Speaker 4:

That's another one where it's, like, living that pain of that of that early start and then having to go kinda start from scratch again, not off a reference design, was super informative on the design we did. So, you know, it it's another one where it's like

Speaker 2:

I I think that this is actually a really important point making, Steve, in that the and I'm sorry to say actually there. The Yeah. The I'm not yeah. I don't mean to sound surprised, but just that when you spend time going down what turns out to be the wrong path, and then you get on the right path, it's hard to lament the time spent on the wrong path because that time gave you conviction about being on the right path. And there are so many of these things where it's like, yeah, we do not actually have a second I mean, we do not wonder.

Speaker 2:

Like, boy, should we have used OpenTitan? Boy, I wonder. It's like, nope. We definitely knew to know. We know that we should not have done that, and we but because we went down that path for long enough to know that that was not gonna be viable.

Speaker 1:

Hey. But didn't they just tape that thing out? I mean, come on.

Speaker 2:

That's right. We could almost be shipped like

Speaker 1:

The the hardware root of trust that we considered very, very seriously, but we're worried that it might not ship in time. And, boy, was that right?

Speaker 2:

Yeah. That's right. What question in the chat. Like, would you still do hardware virtualization, if you're doing it over again? Absolutely, positively.

Speaker 2:

No question. Yeah. There's no I mean, the hardware virtualization. You have to be just in a multi tenant cloud. And, obviously, I think and there was, like, some people I think feel crestfallen because, like, look, I've got some talks from back in the day talking about container based virtualization and lamenting this kind of the fact that we are virtualizing twice in the stack.

Speaker 2:

I still believe that. I still believe all that. But I also, the that hardware virtualized interface has been made first of all, it's been made so fast by the microprocessors. And that's the interface that we've all settled on, for cloud computing. And it's, the I was just because I think the part of what

Speaker 3:

you said too is really important though, which is, like, multi tenant, which you may be like, wait. This isn't a public cloud. This is a private cloud. But, like, with our abilities to scope, like, silos between different parts of your organization, like, you actually do need to treat your internal stuff as though it is multi tenant because, like, it is even if your organization does own it's, like, only things that run by your company, but that doesn't mean your company is not, like, 12 companies with trench coat. Right?

Speaker 3:

So, like, it is important to treat it that way even if, like, you may not otherwise think so.

Speaker 2:

That's it. And I think that that on hardware bird in particular, the reason that that is required is because it does allow for it's it's this very well understood interface where you are allowing for a lot of flexibility inside of that interface. So it's not just, the it's not just Linux. It's so you can run Windows. It's so you can run your own variant of Linux.

Speaker 2:

We've got, like, lots of variants of Linux out there. I'm definitely learning way more about old distros of of Linux than I than I thought I would. So, it's hard revert is important. So, the, and in terms of the overhead, yeah. I mean, yeah, I think that they overhead there's still, there are still challenges certainly in in terms of things like we do not over provision DRAM, we don't over provision CPU.

Speaker 2:

But, the, yeah, block storage performance, happy to do that. People are asking about block storage performance. We can definitely, with and we're we in general, we need to talk a lot more about what we've done from a storage perspective because that's

Speaker 4:

That that feels like one of the follow-up deep dives.

Speaker 2:

Follow-up deep dives for sure. Spend a lot of time on it. It's super important. And so we've got, a lot that we wanna, we Yeah.

Speaker 4:

Of the ones that folks are interested in, one, we already talked about. I'm not gonna drag us back to fans, but just energy efficiency generally. Right. That is one that has a a a it it requires its own kind of deep dive. And then I think there is a networking one.

Speaker 4:

Looks pretty interested in both the hardware and the software side of networking. Yep. It needs to control it end to end and, storage storage right up there.

Speaker 2:

And then another one that that came up a bunch in the hacker news comments, and just in general, is a you know, hey. What about a GPU? And we talked about this above a bunch last time too. The, the and we know there's a lot of energy around GPGPUs for sure right now. At the, at the moment, I would say that we don't believe and if by the way, I know we've got some folks at AdVideo that follow the company.

Speaker 2:

It is our belief that we cannot build oxide value on NVIDIA.

Speaker 4:

As it presently stands.

Speaker 2:

As it presently stands. We would love to be wrong. We would love that it would be great for NVIDIA to say, like, no. No. No.

Speaker 2:

We would love to do something where and and by oxide value, I mean, real transparency and the ability just kinda to to support arbitrary customer problems, to get visibility into power consumption, into firmware, into Security.

Speaker 4:

So they have attestation from the lowest level, you know, first initial instructions all the way up through firmware and and operating system, etcetera. Yep. And it we've we've gotta be consistent on that value because it is security. It's another good one that we need to go deep on, because there's so many different elements of of the system where that pulls through. But it it would be wonderful be wonderful if there was an opportunity to to to build in the same way with the same openness.

Speaker 4:

But today, that's not possible.

Speaker 2:

That that's right. So, I I think that we are looking to the future in a couple different ways. We talked about, that we're we're very bullish on what AMD is doing. But we also know that, like, and, you know, we've talked to plenty of folks and maybe you in the chat as well, you listener, may be of the same, belief. Like, look, it's interesting, but there's a software gap right now between the software, the, you know, and I I think that that's very real.

Speaker 2:

So, and then I think the other thing we're very curious about is, like, there are other things we can go do. AMD and Nvidia are not the only 2 companies in the planet. And there are a bunch of interesting companies that that can do silicon that we can do in some interesting ways, that are, that we would be able to partner with and be able to get real oxide value and kind of an an open source stack. So, on that one, I I can't say contact sales on that, but I guess contact engineering, what do we do on that? And then it's like they

Speaker 4:

Contact us.

Speaker 2:

Exactly. Contact us. Oh, another question was, do you have any regrets about the LPC 5 s 69 given that the number of vulnerabilities that Laura has found,

Speaker 4:

better than the ones you haven't found?

Speaker 1:

And this is the chip that we use for our hardware Root of Trust.

Speaker 2:

Hardware Root of Trust. And I mean, regrets is a really strong word. I I don't know. I don't think so. I think that we I think SecureSilicon is really, really, really hard.

Speaker 2:

And, we, I I do think that we have spent a lot of time with this part, and I think we know it well. And we have been able to build a secure infrastructure on top of it. But it is, you know, it it is a probably a dimension of the product that we are interested to know what else is out there, and we're constantly keeping it here to the rail.

Speaker 3:

Not mad. Just disappointed.

Speaker 2:

Not mad. I'm just disappointed. I am. And and it's it is also it's just it it it's very challenging technically, and there is so much temptation in the industry to have this retreat to security through obscurity. And we I just cannot emphasize this sharply enough.

Speaker 2:

We need a completely open stacks for the secure silicon. And so the the fact that the bootloader is not open is a real problem. And, I mean, Laura would be the first to tell you that, like, actually, if it had been open source, I might not have found this vulnerability, because I went to look for this vulnerability because I popped up Ghidra on this thing because I didn't have the source code. So, this is the and it was the the second vulnerability she found, not the first. So, the it's I would like to see a lot more transparency from our vendors there, and a lot more openness.

Speaker 2:

So no regrets, but, yeah, as you say, Steve, I'm just disappointed. I'm not mad. The, someone's asking about potential cerebrus oxide collaboration. Definitely, that would be, tantalizing cerebrus. We are it's a sibling sibling company, wafer level silicon, really interesting company, and definitely, tantalizing there.

Speaker 2:

I I would say no concrete plans, but, that's something that, folks are interested in.

Speaker 4:

Well, I think I mean, broadly and it comes back to a question that, of course, came up in some of the Hacker News discussions and certainly with reporters, which is like, oh, what are you doing about AI? Because no piece can get written today without talking about AI. Right. And the I mean, our perspective continuing to form is that, first of all, there's this massive general purpose computing market that is in dire need of a solution, a better solution that brings the benefits of cloud computing on premises. That still represents the vast majority of infrastructure that is being deployed and run everywhere outside of the public cloud.

Speaker 4:

And so start there first. And we had customers continually say, like, please please please please please go solve that first. And, actually, wait until this, like, AI wave crest and and then, you know, deliver follow on products that'll help support that. And, you know, I think one of the things that we anticipate is that this rush to smash a bunch of GPUs into systems right now to serve these, like, short term needs are gonna evolve into some real real inefficiency problems around power and cooling and just data center density, where you're gonna have these, you know, quarter filled racks and 8 filled racks. And, so I think as we are thinking about it, it is from a rack level.

Speaker 4:

Yeah. Right. Yes. Thinking about the right balance of hardware acceleration and, in general and just compute and then the power and the thermals. And so we and, you know, same for a question of, like, you know, partnering with any of these other companies.

Speaker 4:

It's like when you start thinking at the rack level, there are some pretty interesting opportunities.

Speaker 2:

There are. And there's some interesting opportunities, but there are also some interesting constraints on working together. And that's what and we I mean, honestly, the way we see it right now is the battle lines are kind of being drawn because NVIDIA at the same time and, again, we'd love to be proven wrong on this, but NVIDIA is also like, actually, I'm eyeing that general purpose CPU, and I'm thinking, why don't I just take the whole thing? And instead of sitting out here on a as a as a peripheral, like, maybe I'm just the computer. And, so I think, you know, the old, like, I'm I'm the captain now meme with with the g p GPU.

Speaker 2:

And I think that that's, that's a proprietary world that we're, we're we're not so into. I mean, I I totally admire NVIDIA's execution, and wish they were a much more transparent company. Make it much easier to to to interoperate. And in in the meantime, I think we really we'll we'll wanna take a different tack, and one that is very much, Steve is just saying, taking that that rack level approach. I did love and Adam had quoted the Adam, you quoted the the the kind of the comments in the chat where someone's like, I really like the fact that it mentioned AI.

Speaker 2:

Yeah. Tell me that's like, yeah. That's because they don't have a GPU. They don't wanna mention AI. We're hoping you don't mention AI, pal, which is not wrong.

Speaker 1:

Not not wrong, but, again, like, there is more computing happening in the universe than just AI, Although, it is consuming most of the power.

Speaker 2:

Absolutely. And we think that, obviously, this stuff is super important and but it's also I mean, I think it's important to get this stuff on package that we that, you know, right now, it should not be a peripheral. It should be you you should have a a chiplet that is an accelerated chiplet, and that should live side by side with a chiplet that is a a general purpose chiplet. And that is, you know, the company, and I know that a couple folks see that. NVIDIA also sees that as the right model.

Speaker 2:

But, you know, that that's kinda what we see. Another question in the chat about, the and, yeah, Wes, exactly, the m I three hundred. What about ARM and risk 5? So risk 5, you know, I love honor risk 5, and I there's a lot that's, great about risk 5. Risk 5 really needs to figure out who it wants to kinda go after.

Speaker 2:

And are you because there's a degree to which they kinda make whatever argument is convenient. And it's like, if you're gonna go after Cortex m, you really need to go after Cortex m, and you really need to provide the debugging support and complete open complete transparency, please, the kind of these proprietary firmware in risk 5 is a real problem. So I want risk 5 to not kind of repeat some of the mistakes that we've seen. Risk 5, we do not need UEFI. Please don't even contemplate it.

Speaker 2:

In terms of ARM, also interesting, but same kinda comment for Ampere. It's like Ampere's slides look a lot less interesting when you compare them to AMD than you do when you compare them to Intel. So, x86 is not just Intel. It's Intel and it's AMD, and, obviously, we're we're very long. AMD and something, are you talking about, wait, risk 5 or sci 5?

Speaker 2:

And that's part of the problem is that that sci 5 is sci 5 has got a lot of proprietary elements. It's been really problematic. And risk 5 is kind of allowed for some of those proprietary elements. So risk 5 is interesting. I I believe it is Laura's Adam, what was Laura's prediction about when we're gonna get a a data center caliber risk 5 part?

Speaker 2:

I

Speaker 1:

think we're gonna come 2 years away now. Yeah. We're getting close.

Speaker 2:

Excited. Yeah. But and

Speaker 1:

I think the real thing is, like, we're gonna listen to customers on that. Like and Okay. There are lots and lots and lots of ways where oxide is weird, too weird. Like, we wanna be normal in some small number of ways and choosing the x8664 ISO is one of the small ways.

Speaker 2:

You sound like Pinocchio who wants to be a real boy. I just wanna be normal. Oxide wants to be normal. Like, I we we do we actually don't wanna be a puppet. We wanna be just a boy.

Speaker 2:

Just a boy. Just a little boy.

Speaker 1:

So, the the the launch press was great. I'm sure you must have loved reading all of it. I mean, it seems like broadly very positive, and great milestone. But, Steve and Brian, as you as you kinda think towards the future, are there milestones that you're looking forward to? And I know that this one was kind of both a starting line more than a finish line.

Speaker 1:

But are are as you look over the horizon, are there are there points that you're excited about getting to in the future?

Speaker 4:

I think the the it isn't a particular date per se, but the thing that I'm excited about is just in these first deployments, we've been able to validate some of the promise for what it's like to take delivery and stand it up and get users going. And being able to hear from customers that this has been a real step function change in terms of how quickly they can go from something arriving to having developers productive has been great. The next thing I'm very much looking forward to is some of the other promised benefits that customers want to go validate, and that is, you know, once I have moved off of the traditional rack and stack approach onto Oxite, was it really the case that I could collapse, you know, 2 racks down to 1? Right. And, is it the case that, you know, now, you know, product teams across these different divisions are much more productive or just happier, or we can recruit and retain folks more?

Speaker 4:

Just some of those, longer term benefits that, that I think these customers were excited to to be part of why Oxide, having those come to fruition, and then hearing them tell that story. I think that getting on the other side of that is, is something that I'm I'm looking for.

Speaker 2:

Yeah. And I think that also be kind of in the same vein. We've got such a great foundation, but it is a foundation. And we've got a lot of stuff now that we can go uniquely do. I mean, I know one of the things that I'm super stoked about is the ability to get, like, unprecedented network visibility.

Speaker 2:

And there's so much stuff we can go build in terms of of of network observability because we control both sides of this, and that can deliver just really eye popping value. And that's gonna be, I think, really and there's not just I feel like in every part of the stack, there is now something we can go do, that it was not really possible before. So I think that we are able to do things that really, really demonstrate the kind of the the integrated, nature of the system, and, that's gonna be that's gonna be really exciting. And I think that, you know, it's gonna be the journey's also gonna be long on all this stuff. We did have There was, at some point, there was a suggestion, maybe from someone not at Oxide that but, like, well, you need, like, an our incredible journey blog post.

Speaker 2:

And it's, like, okay, maybe I'm too online, but, I I no. Our Incredible Journey that's a uag. Sorry. That's, there's, like, an our incredible journey tumbler that is a collection of companies announcing that they're growing out of business. Our journey is not an incredible journey.

Speaker 2:

It is a,

Speaker 1:

Yeah.

Speaker 2:

It it is you know, it's it's kinda funny because I I know I'm that's the same thing, like, you know, when you see that you've got these things that are in that kind of the the the distant horizon. And it takes so long to work there that sometimes you can forget, like, how the you can like, you don't feel necessarily the rush at the end, and you need to, like because you're kind of looking to the next horizon now, and you do need to take a moment and be like, okay. This is like we've actually this is pretty great, and now let's get, what get excited about the about the next horizon. But I remember I mean, Steve, I mean, how long ago was it me sitting in your office at Joyant, just really thinking like, man, what's it gonna feel like when we close that first customer, get that first PO,

Speaker 4:

lifetime ago and not long ago,

Speaker 2:

a lifetime ago, not long ago. And you think that like now that like that milestone now, which felt like it was so indefinitely in the future, pretty far in the rearview mirror at this point, that first PO. And that's that that's, that's pretty great. So, pretty exciting, but a lot to go build. And a a lot to go build, a lot to go do, for sure.

Speaker 2:

So, Adam, does that,

Speaker 1:

This is great. Great. I mean, it last week was was so wild and, great to kinda put a bookend on it, but then, you know, lots of work to do and look into the future. But, you know, we're taking the time to, to really savor this one because it's been a long journey and we got a lot ahead, but, but this is great and it's great to see people responding to it and excited about what we've built.

Speaker 2:

Yeah. I think that's a great note to end on, Adam, because I agree with you. I think the one thing it it was it was really surprising that this is now, like, whatever it is. The 231st most popular story in the history of Hacker News. In the 34,000,000 articles in Hacker News, I and the, you know, you think that that this is, the things that are ahead of us are all involved death or scandal.

Speaker 2:

And, like, that's

Speaker 1:

You mean in terms of stories, not the future of the company. Yes. Agreed.

Speaker 2:

Okay. Sorry. No. No. You're not getting one of lay it down.

Speaker 2:

So we need I I I sorry. We that we need a scandal involving death. I got come on. Don't be thick. Like, we got wait.

Speaker 2:

Like, we we can best it. We we we can we can, best 231. But the, I I think that that's amazing that there's that kind of enthusiasm, because you don't generate that enthusiasm. Like, that that enthusiasm is like that that that comes from a deeper spot, and we feel just really lucky. Lucky that that that folks find inspiration in what we're doing, but we're honestly it it we're the ones that are, really excited by by the amount of of enthusiasm that that's out there.

Speaker 2:

So all of you, you know, the folks that are you know, I I know it's not a home lab variant of oxide, but we really, really appreciate all of your support, and enthusiasm for the company and what we're doing, and, we're excited to be a model for other folks who are willing to take on big, hard problems.

Speaker 1:

Awesome stuff.

Speaker 2:

Alright. On that note, thank you very much, everyone. Really appreciate it. We're gonna have again, we'll do a couple more of these for sure. I know there's some questions in the chat that we didn't answer.

Speaker 2:

I'll I'll kinda hang out a bit, in the chat and answer them, But, thank you very much, Adam. Thank you. Steve, thank you for, for me for combat duty Yeah. In the Hacker News comments. It was it was definitely, it was it was a wild one.

Speaker 2:

So It is.

Speaker 3:

But I do echo what you said about being really appreciative. Like, as much as it's fun to complain about the bad comments, just the fact that there's so many people that are so enthusiastic about what we're doing is just, like, very it's very nice. I appreciate it quite

Speaker 2:

a bit. It it really is. And, you know, when we when we first raised, we said we're gonna have the Internet rooting for us, and, that's been great. It it really, really, really appreciate it. So alright.

Speaker 2:

So, Steve, onto the onto the next. The next. Alright. Thank you very much, everybody, and, thanks again for for all the the the terrific vibes, and, we promise to never ask you to contact sales. Also, please contact sales.