Subscribe
Share
Share
Embed
Chaos Lever examines emerging trends and new technology for the enterprise and beyond. Hosts Ned Bellavance and Chris Hayner examine the tech landscape through a skeptical lens based on over 40 combined years in the industry. Are we all doomed? Yes. Will the apocalypse be streamed on TikTok? Probably. Does Joni still love Chachi? Decidedly not.
[00:00:03.15] - Ned
When I need help with something, I should probably reach out to the people who are capable of influencing a decision and rendering assistance.
[00:00:14.20] - Chris
I mean, counterpoint, instead of that, do nothing and slowly wither away in the dark, in the corner, all alone into a nonsensical husk.
[00:00:27.09] - Ned
So what I've been doing for the last year. Hello, alleged human, and welcome to the Chaos Lover Podcast. My name is Ned, and I'm definitely not a robot. I'm a real human person with a husk in a corner, and it's very dark, and I'm sad and crying. And for whatever reason, Bob's Big Boy is there, and it is, it is hella creepy. With me is Chris, who is also in the corner.
[00:00:57.13] - Chris
That is me in the corner. That's me in the spotlight.
[00:01:03.06] - Ned
I would ask you if you're losing your religion, but I think that ship has sailed.
[00:01:08.06] - Chris
I don't want to get political.
[00:01:12.06] - Ned
Oh, yeah. That is the third rail, isn't it?
[00:01:16.11] - Chris
REM? Yeah.
[00:01:18.06] - Ned
A lot of people—
[00:01:19.03] - Chris
we all know they should have retired when their drummer left, like they promised they were going to do. And unfortunately, they didn't, and they never left.
[00:01:28.22] - Ned
I mean, in fairness, have they actually made a good song in the last 10 years?
[00:01:35.11] - Chris
Has there been a good song in the last 10 years?
[00:01:39.14] - Ned
Well, I can tell you who hasn't made a good song. It's R.E.M. or U2. Or any other very shortened band name.
[00:01:48.14] - Chris
Yeah, right. You don't even give us the respect of using full words. I'm not listening anymore.
[00:01:55.05] - Ned
I wrote U2 off as soon as they pushed an album to every single person's iDevice.
[00:02:02.06] - Chris
You don't even have an iDevice and you were appropriately outraged.
[00:02:05.12] - Ned
Could you imagine how angry I would be if I actually had? An iPhone or an— I think I probably did have an iPod Shuffle, but I'm pretty sure nothing could be pushed onto that.
[00:02:17.19] - Chris
Or maybe a Nano? Those were cool. Because they were—
[00:02:21.17] - Ned
yeah. Was that the Nano?
[00:02:23.12] - Chris
Yeah. And they were super easy to make waterproof.
[00:02:25.21] - Ned
Yes. That's why I had it.
[00:02:28.02] - Chris
Yeah.
[00:02:29.20] - Ned
Yeah. I actually attempted swimming with it a few times. And while the Nano worked fine, the headphones
[00:02:38.18] - Chris
Less fine.
[00:02:40.10] - Ned
It's hard to keep them in your ears and it's just, just don't do it.
[00:02:45.21] - Chris
Well, they make headphones specifically for swimmers now.
[00:02:49.07] - Ned
Yeah.
[00:02:50.14] - Chris
But we could also remember that we're not fucking dolphins and swimming is stupid.
[00:02:55.22] - Ned
Exactly. Let's talk about some tech garbage.
[00:03:03.07] - Chris
Privacy-focused email and VPN company makes an awkward as fuck pivot into AI development, and I don't know how to feel. No, I'm actually— I, I hate it. So Proton, the Switzerland-based company behind ProtonMail, ProtonVPN, ProtonDrive, and a few other privacy tools that don't start with Proton.
[00:03:27.17] - Ned
Really?
[00:03:28.11] - Chris
had an internal email leak last week. In it, leadership told people at the company to expect to re-interview for their own jobs. The focus of the interview was to discuss productivity, primarily productivity from employee use of AI. Employees were to be interviewed by people outside their own leadership chain who, quote, already demonstrate strong AI capability, unquote. AKA, They're going to be interviewed by people who are already all in on this initiative. Proton hastened to say that this was not a layoff engine, but they also said these re-interviews are consequential. Interviewers will document their assessments, which will inform the broader review of our organization and talent. That review may lead to change is— it's going to be a long day, kid.
[00:04:16.15] - Ned
Yep.
[00:04:18.01] - Chris
That review may lead to changes in roles, responsibilities, and continued employment. So what? The email naturally leaked, causing a furor. Proton then released the email in full on their official Reddit sub, hoping to make things better, instead causing a furor-er. Then the CEO of Proton, Andy Yen, had some previous bad behavior leak. He's apparently all in on AI, on producing more and more and more and more and more, and doesn't seem to give a shit what anyone thinks about it. On the company's Slack back in August, an engineer suggested pen testing new code before pushing it to production. Crazy idea.
[00:05:05.21] - Ned
Hmm.
[00:05:07.18] - Chris
So this employee is urging caution and security-minded thinking about a push to prod. And Andy Yen's response was, quote, that's bullshit and loser thinking. Of course it's possible to be fast and secure. Stop thinking always of how to be slower and focus more on how to be faster, unquote. In a follow-up message, Yen reiterated his point by saying, quote, only losers think like that. His ultimate message for the team was to stop lovingly handcraft all your code and use AI more, unquote.
[00:05:42.17] - Ned
All right.
[00:05:44.13] - Chris
This, to my mind, flies in the face of Proton's stated mission of being secure and private. Ruthlessly smashing everything that stands in the way of AI velocity, to my mind, is the kind of dumb shit Twitter does. And you know how well Twitter's been working these days.
[00:06:00.19] - Ned
True.
[00:06:01.19] - Chris
Responsible companies should be a bit more circumspect
[00:06:05.23] - Ned
and
[00:06:07.18] - Chris
You know, Andy Yen seems like a giant dickhead, and I am sick of every IT company on earth being run by obnoxious giant dickheads.
[00:06:15.19] - Ned
Indeed.
[00:06:17.03] - Chris
The situation does not make me feel good about a company who in years past have had a great reputation and were my first recommendation to people who asked about fleeing Google. Update, prior to recording this episode, but after writing everything you just heard, Andy Yen decided to cancel the re-interviews. He claims that the furor meant that the interviews were unnecessary and he had already gotten the message across. So that's fun. Employees at Proton said the AI usage is just going to be, you know, part of ongoing performance reviews instead. So I take it back, that's not fun.
[00:06:57.19] - Ned
Nope.
[00:06:59.10] - Chris
Yen, ever the coward, did not actually apologize for his shitty behavior. He just said it was a Communication failure. He also said, going forward, I'm still going to push on his AI initiatives. Safe to say that part of the message was probably received. He felt no compunction, however, to walk back calling his own engineer a loser for wanting to pen test new code. Of course, that would be un-CEO-like in 2026.
[00:07:29.15] - Ned
Never admit fault.
[00:07:31.15] - Chris
God Almighty, every word I read about this prick makes me hate him more. At least the CTO at Proton, Bart Butler, could be direct. Quote, we fucked up. We know we fucked up. Yes, Bert, or Bart, or Brat. I don't even care about your name. You're part of the problem, but good on you at least for being able to read the writing on the wall.
[00:08:00.21] - Ned
Poems in space. Let me try that again. Poems in space.
[00:08:10.17] - Chris
Much better.
[00:08:12.04] - Ned
Chris, space is a cold, uncaring, almost vacuum filled with cosmic radiation and a total lack of wah-wahs. I mean, seriously, if you really want people to go into space, You've got to at least establish a Wawa. I'd be okay with a Sheetz or a Buc-ee's, but as we all know, you gotta have a Wawa even in the unforgiving blackness of space.
[00:08:42.00] - Chris
Also, fuck Buc-ee's. If you know, you know.
[00:08:45.18] - Ned
What about data centers though? We sure do hate data centers. They don't even have a meatball sub. Or the seasonal gobbler Shorty? Why not send these noisy, energy-hungry disasters into space? That's what Google is thinking, and they successfully launched their first orbital AI data center satellite called Suncatcher in early October. Since then, they have communicated with the satellite and started spinning up its hardware. The satellite is powered by both an internal source and solar panels. Inside the satellite are 4 Google Trillium TPUs, with a single TPU operating at a time. Last week, the team at Google Research were able to load a Gemma 4 model onto the TPU and send it a prompt. What was this prompt? A request for the model to create a haiku about being in space. The model returned the following. You'll need to add some pan flute music. Thank you. Blue marble below, data streams in silent grace, starlight feeds my core. Aside from that being a lie because it has an internal power source, the haiku also ended with an emoji, a star emoji. And despite being silent, I believe that adds an extra syllable to the poem, making it no longer a haiku, you have failed.
[00:10:15.18] - Ned
Clearly this whole project is a failure and Google should scrap any future plans for data centers in space. Assuming Google does not heed my advice, the next series of tests will be around heat dispersion using a specially crafted heat pipe and radiator system and the system's resilience in the face of elevated cosmic radiation. The Register published some rough estimates on the cost of this haiku-butchering satellite, and it comes out to $3.5 million. That is a lot of money to spend on bad poetry. Future missions are set to launch 2 satellites for double the cost and test intersatellite communications using free-space optical links. That means lasers. I like that part.
[00:11:05.01] - Chris
We like lasers. Yeah, we like lasers.
[00:11:07.19] - Ned
From an engineering standpoint, I have to admit this stuff is pretty cool, but I have this nagging thought in the back of my brain that running data centers in space is wildly uneconomical. Google estimates that launch costs would need to drop from their current price of $7,000 per kilogram to $200 per kilogram. To make the program economically viable. Now they are projecting that SpaceX AISI could hit that price by 2035, which I will say is just past wishful thinking and entering dangerously deluded territory. One might even wishcast a Wawa on the moon. If $200 per kilogram launch cost is possible, Why not a Sizzli on the Sea of Tranquility?
[00:12:03.10] - Chris
I don't think the world needs a new GitHub, says, um, GitHub. Sometimes I feel like all we really need is the headline. I mean, honestly, what else do we think GitHub is going to say? They're just going to like roll up and be like, Well, we actually welcome the competition because we believe GitHub is superior in the market and will win customers on its merits alone. We don't need to rely on things like market pressure, outdated name recognition or reputation, or licensing fuckery. We're just better.
[00:12:39.22] - Ned
Adorable.
[00:12:41.17] - Chris
Back to the real world. Gone are the days of just deciding between GitHub and GitLab. And I guess a zip file on your hard drive? No, sir. Thanks to, you guessed it, AI, the market for what should be a relatively simple technology is exploding. Cursor, recently and unfortunately acquired by SpaceX in a deal that finalized in August, released their own code repo product called Origin. The name's actually a lot longer than that, but we're sticking with Origin.
[00:13:12.14] - Ned
All right.
[00:13:13.06] - Chris
The timing of the Origin release could not have been better for Cursor. As that was the same day that GitHub went down for almost 8 hours. Recently, Thomas Domcke, former CEO of GitHub, released a new product called Intire, which is a stupid name too, but touts to be a GitHub competitor that's built for modern code repo workloads. And by that, of course, he means for fucking AI.
[00:13:42.06] - Ned
Yep.
[00:13:43.16] - Chris
The problem with AI coding is that autonomous agents are logging changes and pull requests at just a ridiculous volume. Back in June, GitHub's commit volume was approximately 1.4 billion transactions in a month. Previously, Google— Google— previously GitHub had been doing 1 billion commits in a year. So that's a big increase. But in September, the number was 7.4 billion. So a bit of an expansion.
[00:14:20.08] - Ned
Slight one.
[00:14:22.04] - Chris
And the GitHub architecture, it seems, just cannot grow that fast. According to GitHub's own numbers, by the way, which is a website, uh, what is it called again? It's called something. They saw devs pushing code changes to GitHub 17.9 billion times in the last 12 months, more than 2024 and 2025 combined. Safe to say this pace is not exactly slowing down.
[00:14:50.11] - Ned
Not even a little.
[00:14:52.18] - Chris
At the Universe conference this week, GitHub's leadership basically threw out a just trust me, bro, and said that GitHub is working hard And evolving in the agentic area. I hate words. In the agentic era. What sets GitHub apart, they said, is that they can handle so many, very many projects and so many, very many developers. Except, you know, that time they knocked like half a percentage point off their uptime stats for the quarter in a single afternoon.
[00:15:27.08] - Ned
Yep.
[00:15:29.17] - Chris
We'll have to see what happens. I mean, there is no way that GitHub is going to go away anytime soon. It's got an absurd number of customers. It's like 250 million. It has the backing of Microsoft. And to be fair, there is a lot more to GitHub than just the repo functionality. Developer people claim that that's important.
[00:15:53.02] - Ned
Yeah.
[00:15:54.12] - Chris
But GitHub is also like the 800-pound gorilla in a room full of hungry barracudas. Wait, what? That's a terrible analogy.
[00:16:04.16] - Ned
Yeah.
[00:16:05.01] - Chris
Let me try again. GitHub is like a battleship in a sea of destroyers.
[00:16:12.20] - Ned
Better.
[00:16:14.12] - Chris
GitHub is probably impossible to sink, but it ain't exactly nimble. And lest we forget, Germany thought that the Bismarck was unsinkable. And we all know how that worked out for them. Now, don't we?
[00:16:29.16] - Ned
You have sunk my beautiful ship.
[00:16:33.23] - Chris
That— what country was that?
[00:16:35.23] - Ned
That was a quote from Bill and Ted's Bogus Journey.
[00:16:42.16] - Chris
I know where the quote came from. I just want to know what you did to it and why it deserved that accent.
[00:16:47.04] - Ned
That is how death says it. It wasn't me. Talk to the actor. Tape is dead. Long live tape. This is my annual act of contrition regarding tape as a backup and archival target. Almost 10 years ago, dear God, you know, 10 years, I said tape is dead. And every year since then, sales of tape have grown and the available capacity for a single tape cartridge has increased. Now, I still hate tape with the power of a million exploding suns, but it appears I am in the minority. Fujifilm announced last week that they will be investing $63 million to increase their tape production capacity fivefold. Why the big increase? It's 2026, so you get one guess.
[00:17:49.01] - Chris
Wawa.
[00:17:50.03] - Ned
Oh God, I wish. It's AI, because of course it fucking is. Well, it's not just AI, but, well, I mean, it is kind of still AI. It's just like tertiary effects. The increasing sophistication and frequency of cyberattacks and ransomware ransomware have made tape an attractive place to store your most important data. Tape can sit offline, consuming no power and completely air-gapped from whatever horrors Jerry from accounting has unleashed on your LAN. Fucking Jerry. I was Jerrying it up.
[00:18:25.02] - Chris
He's probably a Cowboys fan and you know how those people are.
[00:18:28.19] - Ned
They don't even like meatball shorties. Now, where did this increasing sophistication and frequency of cyberattacks come from? It's AI. So yeah, there's that. Tape is also behind many of the archival solutions offered by cloud hyperscalers. If you ever wondered why Amazon Glacier takes so long to bring data back online, that's because it's restoring it from tape, a painful process that is linear in nature And horrifying in practice. Fujifilm will be focused on building out additional capacity to produce LTO-10 generation tapes. LTO-10 is the current generation and can store up to 40 terabytes native or 100 terabytes of compressed data per cartridge. The current LTO roadmap extends out to LTO-14 with a capacity of almost a petabyte per cartridge compressed, although no dates have been proposed for when LTO-14 will drop. Sales for LTO-10 tapes surged by 57% in the first half of 2026, so it's no surprise that Fujifilm and others are ramping up to avoid the issues plagued by RAM and SSD vendors in the last 12 months. It's almost like— And stay with me here, AI is making everything worse and more expensive, even tape.
[00:19:59.08] - Chris
Apple announces the most powerful, feature-rich MacBook Pro in their history with touch-enabled display almost as an afterthought. Okay, so first the specs: thinner, lighter design, 14 or 16-inch models, M5 Pro and M5 Max chip options, AI capabilities, because of course, a claimed 24-hour battery life, a brand new and apparently kick-ass dual-layer OLED display. Oh, and if you touch the screen, um, for the first time ever for Apple on a laptop device, Stuff happens. All for the low, low price of, I promise you really don't want to know.
[00:20:50.22] - Ned
Not even a little.
[00:20:53.11] - Chris
So the touchscreen is what got all the press, right? And I mean, fair. Touchscreen on a computer is something that Apple has famously resisted. Steve Jobs, in a clip that's going viral all over again, once said that the ergonomics made a touchscreen on a computer, aka a non-tablet, A nun— a dumb idea. I almost said non-starter and dumb idea, and I said nun idea. You know how nuns like to do the TikToks.
[00:21:24.05] - Ned
It's gruel.
[00:21:27.07] - Chris
Now, it wasn't that Apple couldn't make a touchscreen. It's just they thought it was pointless, and they just haven't made one. Important difference. And besides, Jobs, well, Jobs routinely talked shit about product lines and software features that he said were dumb or unnecessary and turned out to be markets that Apple eventually decided to jump into with both feet. The touchscreen is kind of limited in what it can do. You cannot use the Apple Pencil on it, for example. It's only passive styluses. Styluses. Stylus. You can click, you can hold and click, you can expand to zoom. It's basically mouse gestures built into your hand. And for the moment, the touchscreen is only available on that one model, which happens to be the highest-priced laptop Apple makes. Safe to say that touch is going to be a bit of a slow growth cycle. And maybe a slow growth cycle for new technology or new features is not the worst idea in the world. After all, they forced their dumb folding phone into the market way too early, and all those things seem to do is break at the hinge point. What a shock. Apple's own release makes it pretty clear that while they're talking about the touch aspect of it, it's kind of an afterthought.
[00:23:00.09] - Chris
New CEO John Ternus said, quote, We needed to get to a point where adding touch didn't make the product lesser in any way. It doesn't make it heavier, it doesn't make it thicker. That came together with the OLED screen, the best display we've ever put into a Mac. All of these things, and oh yeah, it has touch as well, unquote. I mean, 2 points for honesty, right? And All right, fine. I know you really want to know. The cheapest one of these fuckers that you can buy that has the touchscreen starts at about $2,900. Emphasis on starts. I personally don't need anything close to that much horsepower in a laptop. Truth be told, I am still doing just fine with my M1. I haven't bought something from the MacBook Pro line since I was in grad school. And honestly, even then, it was probably overkill. So who knows? Maybe this is a proof of concept, the touchscreen thing. Maybe it'll grow over time, or maybe they'll just leave it out to wither and die like they did with the Touch Bar. Oh, last thing about this release, and this is a major thing for some people.
[00:24:19.02] - Chris
The new laptop has the glowing Apple on the back. Just like it did back in the early 2010s. Will some Apple fanboys kick out 3 large or more just for a glowy Apple thing on the back of their laptop case when this goes GA on Friday? I think we both know that they will.
[00:24:41.14] - Ned
They shall. AI is eating cloud too. I really am on an AI makes everything worse kick today, and that was not intentional. Well, okay, maybe it was like a little intentional. And for that, I apologize. To no one. Hahaha. Let me ask you a question. Chris, have you recently tried to provision a virtual machine on AWS or Azure and the request immediately failed or hung for like 30 minutes before telling you it had insufficient capacity for your request?
[00:25:17.19] - Chris
Let's go with yes.
[00:25:19.09] - Ned
You're a liar. Lying liar. Well, if you have encountered this issue, you are not alone. Amazon CEO Andy Jassy recently said that capacity on EC2 was sold out for 2026, mostly committed for 2027, and already selling into 2028. In this case, Jassy is talking about the block of EC2 instances AWS sells as reservations. There is still some block set aside for on-demand instances, though how big that block is, he did not say. A reservation on AWS or Azure allows you to buy compute capacity upfront for a reduced cost and should, in theory, protect you against a lack of capacity on the cloud provider. Should being the operative word here. Before signing any agreement on reservations, you should check and see if that capacity is guaranteed and if there are any caveats about placement or regions. If you need 20 instances tomorrow in the US East 2 region and the only thing they'll give you is someplace in Canada, that's not the same. Microsoft also chimed in to say that their capacity woes on storage, compute, and GPUs will continue through the rest of 2026 and into 2027. With all that being said, if you're planning on deploying large fleets of compute, especially something with GPUs in the near future, reservations are not just suggested, but basically required and probably already sold out.
[00:26:59.23] - Ned
You'll also need to wear a blazer or a suit jacket. You Godless flannel-wearing hippie. If you're trying to procure on-demand instances, the chances of a failure go up dramatically. Personally, I was leading a training for Terraform on AWS 2 weeks ago, and several availability zones simply had no T2 or T3 capacity. We had to change regions to continue doing the labs. When taken as a whole, The attractiveness of running your applications on-prem begins to actually increase. Sure, you cannot allocate capacity on demand, but you also know exactly how much capacity you have. And if Jerry from accounting is hogging resources with his weird finance app that is definitely a cover for a Minecraft server, you can simply boot his ass. Hey, thanks for listening or something. I guess you found it worthwhile enough if you made it all the way to the end. So congratulations to you, friend. You accomplished something today. Now you can go sit on the couch, launch your own AI satellite into orbit, and ask it to write you a sonnet. You've earned it. You can find more about the show by visiting our website, chaoslever.com, where you'll find show notes, blog posts, and general tomfoolery.
[00:28:19.20] - Ned
We'll be back next week to see what fresh hell is upon us. Ta-ta for now. You're muted. Did you know you're muted? Because you're muted.
[00:28:32.09] - Chris
I figure that's best for everyone.