Subscribe
Share
Share
Embed
AI First with Adam and Andy: Inspiring Business Leaders to Make AI First Moves is a dynamic podcast focused on the unprecedented potential of AI and how business leaders can harness it to transform their companies. Each episode dives into real-world examples of AI deployments, the "holy shit" moments where AI changes everything, and the steps leaders need to take to stay ahead. It’s bold, actionable, and emphasizes the exponential acceleration of AI, inspiring CEOs to make AI-first moves before they fall behind.
Andy Sack (00:01.351)
This is AI First with Adam and Andy, the show that takes you straight to the front lines of AI innovation and business. I'm Andy Sack and alongside my co-host Adam Brotman. Each episode, we bring you candid conversations with business leaders, transforming their businesses with AI. No fluff, just real talk and actionable use cases and insights for you. Today's a special episode. It's just Adam and I, and we're going to...
do a shorter episode than normal with no guest. Adam, glad to have you, glad we could make the recording work.
Adam (00:36.267)
Yeah, I like these little brief special episodes we do with just you and I, everyone's know.
Andy Sack (00:42.959)
Yeah, so you got to be in your bonnet this weekend about the MIT study. Do you want to tee up the conversation that you want to have with me and the audience?
Adam (00:51.373)
I, yeah, so it's interesting. There was a headline that went around last week that I think it was Fortune magazine or fortune.com and they were citing a paper that had come out like the week before from MIT about, know, the headline from the fortune article was 95 % of corporate AI projects.
have been failing or something like that, according to MIT study. And it was like, I did the sideways dog look like a lot of us do, which is like, is, I mean, I'm sure a lot of these pilots fail, but I was shocked at the headline. I, I, but it was too late. I did my homework, but by the way, it was sort of too late, like to rebut it, not that we would rebut anything, but like,
Andy Sack (01:38.225)
You did your homework.
Adam (01:45.26)
But it made its way all the way to like CNBC and there were like stock billions of dollars of stock valuation that came down because of people were wondering like, Oh my God, is that true? I think there's such a tension around whether or not there's a bubble, which there is on some level around AI spending, AI stocks, AI valuations. so like it made its round and I, yeah, over the weekend, I didn't think much about it other than like,
That's weird. That's a weird headline. You know, cause you and I live in this world and we're like, I don't know how anyone could say that. Not because we were saying 95 % of them succeed. was just such an out of left field number and study. over the weekend, I was like, you know, I'm going to read this MIT paper. Uh, and when I did, it got me all fired up because mostly in a good way, like as in, wasn't like this, it wasn't like the paper itself was terrible.
Andy Sack (02:41.426)
you
Adam (02:44.14)
I think the headline was clickbaity and not great, the actual study itself was fascinating because they went in and they talked to 52 companies and they read 300 different reports that were on the internet that weren't even like conversations they had had. They did some surveys of 153 other surveys and they came to the conclusion that a lot of companies were trying
things like trying wrap apps effectively trying like the, either to build their own wrap app or to use a wrap app and only 5 % of them crossed two hurdles that they called success. had to the wrap app or the internal project had to go from pilot to production and they had to answer in the affirmative. Yes, this had immediate P and L benefits immediate. So they had both things had to happen. And I was like, that's a
Again, that's also a weird way to define success is that you had to see an immediate P and L impact. But putting that aside, when I dug in further, it was mostly the articles and mostly about how these. These enterprises that were trying these pilots were realizing that chat GPT on its face was actually easier and more effective than some wrapper app that they were being asked to use by some vendor or whatever. And, and I was really struck by that because
We've been saying for months, almost years now that, that the, that it's actually the individual user, maybe even the micro use cases, the micro solutions that were like so powerful and people were skipping over those and wanting to go to like, tell me what the shiny app is that I can use to immediately save money or boost my, revenue.
Andy Sack (04:24.795)
micro use cases.
Adam (04:41.491)
And we've been saying that for a long time, like, don't do that. Don't, don't try to, don't just try to skip there because if you do that, you'll miss the ability to even evaluate what is the right use case, what is the right app to build and whether or not you should use it or not. And I, and here was this study that was actually saying that in a, in a kind of a convoluted way, they were saying, well, geez, you know, it's hard for these, it's hard for these wrapper apps and these sort of custom apps, call them to be better than what the
the generic, they called it generic chat GBT could do on its own. And I, just thought that was interesting. Like somehow that turned into this cockamamie headline, but, but, but in the, actually, when I read this, the paper, thought, I think the people that wrote this paper would totally agree with us. And we would actually get along and be very aligned on how we think about these things. It just got twisted. And I thought it was interesting to, to, to talk about given how viral the headline.
Andy Sack (05:39.325)
Yeah, I mean, clearly it had a good headline and the virality of it speaks for itself. think that the, make a really good point. mean, this whole notion that 95 % of them fail if they don't immediately produce revenue in the first 90 days, like, um, a, your point about micro use cases B clearly AI is having impact on automation of back office. There was another research article about that.
I think it is going to help companies accelerate top line as well, i.e. revenue. But I'm sure that's harder. So I'm glad that you got your knickers in enough of a twist to write an article. So either you'll be posting that on LinkedIn and it was enough to warrant a quick job. While we're talking and we have a couple more minutes, any comments about...
Adam (06:22.206)
Yeah.
Andy Sack (06:36.977)
chat GBT five or other topic that you want to just highlight for the audience.
Adam (06:41.49)
Yeah, I thought chat to be T five is another one. It's, it's interesting. It's like we're entering into this era of, I don't know if it's going to, yeah, the trough of disillusionment and that I think that the MIT article and even a lot of people's response to check to be five. And I actually thought it was interesting. Ezra Klein wrote a blog post this weekend on the New York times talking about Chachi BD five saying he can't believe all these people are so negative on it when it's actually.
Andy Sack (06:49.841)
Disillusionment.
Adam (07:11.54)
quite impressive in so many ways. and, and I actually think that that it's worth hovering over that for a second. Like there were, we're getting to a place now where there's people tend to be, it's almost like they're politicizing like AI when, meaning like, I'm anti AI or it's not a big deal. Or it's, know, if it's not there, in other words, they're over hyping and then, and then overreacting to the negative both.
Andy Sack (07:14.013)
Yeah.
Adam (07:40.924)
And in reality, it's, I think Bill Gates said to us, and he said it many times that people are probably over hyping, what AI can do in the immediate term and under hyping what it can do. And I think Eric Schmidt said that as well. And so basically like, w what's interesting about, I think chat to be divided is like, is Rorschach or whatever that word is the Rorschach test or I've like, depend. It depends on like.
Andy Sack (08:06.471)
Yeah.
Adam (08:08.745)
where you're coming from. If you were expecting it to be AGI or whatever, you're going to be underwhelmed. And they also had some bumpiness in their rollout that actually made a bunch of people that were sort of hoping it would not do well, you know, talk it down. But in reality, it's a very, it's an incredibly powerful model. Once again, it set benchmarks and it's actually bringing a lot of people to using reasoning models in a way they weren't using them before. It's forcing
It's forcing people to, you know, get better at their prompting in some ways, because it's so good when you actually give it a really thorough prompt. so, in all those ways, I think, I think we'll look back at GPT-5 not as the Death Star release from OpenAI, but at which I think some people were expecting cause Sam was posting these Death Star memes, but actually pretty fricking good release in the end of the day. And
You know, they march on to do a lot of amazing things. And so does Gemini with, you know, via Google's Gemini. And, and I think, you know, to me, GPT-5 is another example of, I think we're becoming somewhat numb to, we're like boiling frogs and we're becoming numb to how fast this stuff is happening. know, reasoning models came out, what, six months ago, seven months ago.
Like came out, I'm talking about 01. Like, so it's, we're, think, I think GPT-5 was kind of a good time for us all to realize like the models are not plateauing. They are sort of bunching around one another. So Gemini, GPT-5, you know, Claude's new models, Grok's caught up amazingly. And like, you're getting these frontier models that are sort of bunching. And even though one's a little ahead.
Andy Sack (09:37.148)
Yeah.
Adam (10:06.402)
there's a lot of interesting implications to that, but I, I, that, that was my, my, my read on GPT five was like, it was, it's kind of a sign of the times about this, like trough it's people are starting to like overreact in both directions. I think now, and, you're missing the point, which is this is. Fucking amazing technology and it's.
And it makes mistakes and sometimes it's not amazing, but overall net net, like I'm still in awe at how powerful this tech is. And, and we, we shouldn't take our eye off the ball that this is something that if you're not really leaning into your, you're probably, you know, you need to start leaning into this a bit and not with hype, but with realizing how powerful this tech is. were your thoughts on GPT-5?
Andy Sack (10:59.005)
I mean, I don't really have much to add. plus one, like I thought JGPT-5 was a really solid model that had a bumpy first couple of days. And I actually found the output much like Ezra Klein to be better. I continue to use JGPT and Gemini as my primary two models. And that was impressed with JGPT-5.
Adam (11:25.128)
I did talk to, before we wrap up on this little mini episode, did, I talked to a friend of mine who runs an engineering team and has always been a software manager. And she was kind of a little bit late on AI and coding for AI, but she got into it a few months ago and she came running up to me after GPT-5 came over and said, Oh my God, it's I now, this is my go-to, not Claude. And Claude is like the bomb when it comes to the goat.
Andy Sack (11:54.181)
Yeah, I mean, think that that was, that there was, definitely made a play. They saw that how much revenue and progress Claude was making with that audience.
Adam (12:05.125)
Yeah. And they inspired, they spent a lot of their demo for the release of GBB vibe on vibe coding. Yeah.
Andy Sack (12:14.013)
All right. Well, with that, thank you all for listening. We'll have more of these short episodes going forward. Adam, thank you for your time on this late afternoon. This is AI First with Adam and Andy, the show that takes you, whoop, that's the opening, pause that. Thank you all for listening to AI First with Adam and Andy. For more resources on how to become AI First, visit forum3.com.
You can download case studies, research briefings, executive summaries, and join our email list there. There's lots of amazing material on our website for leaders looking to expand their knowledge. We truly believe you can't over-invest in your AI learning. Onward.