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 (00:00)
And we still firmly stand by the prediction. We were just early.
Adam Brotman (00:05)
we were right that it's a big topic and the things were going to be significantly disrupted in some ways.
Andy Sack (00:11)
So I think we overstated the prediction
Adam Brotman (00:14)
so I think it was a good prediction in a lot of ways
Andy Sack (00:18)
was a great buzzword for the year, but I'm going to say that we got that one wrong.
Adam Brotman (00:21)
like the theme of this whole review, we got that a little bit right and a little bit wrong. so it's interesting,
Andy Sack (00:27)
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, actionable use cases and insights for you.
Welcome everyone to the 2025 year in review episode. Adam, we've got an exciting plan for today's episode. We're gonna review predictions that we made last year at about this end of 2024 for 2025 and talk about what actually happened in the year and what we got right and wrong in our predictions.
Adam Brotman (01:24)
Great.
Andy Sack (01:25)
So I think it's first, let's dive right in. Our first prediction was that AI search would redefine shopping and Google's dominance in search would finally be pierced. What do you think?
Adam Brotman (01:40)
It definitely was not pierced. but there, was a pretty epic year when it came to AI search. mean, Google themselves, the topic of the prediction, they introduced AI overviews in a very significant way powered by Gemini. And I really think, you know, being willing to sort of disrupt themselves and yet
by the last earnings report, they were showing no slowdown in search and search revenues. I think there's been a big slowdown in traffic coming organically from clicks from Google in the year. So I think if you're on the publishing side, you felt a big change. So I think we were right to say we were right that it's a big topic and the things were going to be significantly disrupted in some ways.
At the end of 2025, Google is navigating that kind of disruption very well. their search revenue is up, their AI overviews are in play, they've launched Gemini 3, which is powering it all, and they just don't feel like they've been pierced in any way yet in that regard.
Andy Sack (02:48)
Yeah, I think our prediction was...
it was a good prediction for a prediction episode, but it was too simplistic. think clearly, we definitely moved towards a way from simple Google search to more AI search. So the user behavior, the expectation by consumers definitely changed in 2025. I I think Google's dominance was pierced, but it didn't totally fall off a cliff. there was a
real rise and perplexity a great replacement for Google search. think perplexity continues to do well. as we're as of the time of this recording, Gemini three just got launched and you know, the the the AI ecosphere is a buzz with that. So I think we overstated the prediction
got it right on the consumer side. I think your point about the publishers is spot on. think if you're on the publisher side, you're
definitely seeing a fall in traffic. don't believe Google has an incentive to have their monopoly position and search traffic not fall. So I questioned some of that accounting, but I think we overstated the Prediction number two, Elon Musk's power play.
Elon Musk would attempt a major power play in AI, especially when considering open AI.
Adam?
Adam Brotman (04:08)
I mean, that one, I mean, that one's a fascinating topic because we, you were, I think you came up with this prediction that you thought he would really go after OpenAI and try and stop them from being a for-profit corporation, thwart their IPO plans and really go after them. I, know, in a lot of ways you were right. Like, I mean, he didn't, they ended up going to a for-profit. They ended up getting that done. But I think,
I think there's like carve outs for the lawsuit with Elon Musk. that lawsuit is not over. And I don't know what that really means practically, but Elon Musk did a lot to, you know, push back on OpenAI for the reasons he mentioned. It was a theme throughout the entire year. It didn't like die down. And if anything, at the same time, tangentially, XAI
know, bot Twitter, Grok has been on fire. It's really come along and has become a significantly powerful frontier model. And so I think it was a good prediction in a lot of ways
like that. know, it didn't thwart open AI directly in a way that maybe the prediction was meant to be sort of provocative, but I think the spirit of it was right on.
Andy Sack (05:26)
Well, at the time when we made the prediction, was Donald Trump was coming in office. Elon had just helped him win the election. it was, you we saw for the first part of 2025 really the height of Elon Musk's influence in political spheres as well as business. And so I think the prediction, and then he did in fact sue
He tried to, he did use that influence and his frustration and sued OpenAI in an attempt to slow them down and take some retribution for his perceived.
the misconducts by Sam Altman and OpenAI from the nonprofit origins. So I think from that perspective, I think the prediction was largely accurate. He did try to use his influence and he did make a power play. I think that's still playing out in the courts and it remains to be seen whether OpenAI is gonna be able to go public. They're clearly on path, they've now transitioned and they're dealing with that lawsuit. We just had the Ilya
Adam Brotman (06:15)
Yeah.
Andy Sack (06:33)
Sutskiver deposition that got released about two months ago, has provided some fun entertaining moments. So I it was a good prediction.
Adam Brotman (06:43)
Yeah, and I think that it's, mean, while the complexities of the lawsuit itself, but the nonprofit and what they did and the origins and all and IPO put that all aside, like, it's just been interesting to see X AI and Elon, you know, it's just remarkable the progress. ⁓
Andy Sack (07:02)
Yeah, continue
to execute with big visions and real moves. I love this next prediction. I think this one was mine. I mean, I'll take credit for it. the year of Apple AI, 2025 would be the year Apple finally showed up in AI in a big way. What do you think, Adam?
Adam Brotman (07:06)
Yeah. Yeah.
If it wasn't
yours, I'll take the blame as well. Cause I, I felt the same way at the time. So, uh, you know, we both, we both thought it would happen in 2025. It didn't, uh, but they're rumored and reported to have done a deal with Google, uh, to license Gemini, I think Gemini three pro or the like, uh, into Apple intelligence, iOS, Siri, et cetera.
It's all, don't really know. So you gotta believe 2026 will be the year. I will say they, in 2025, understanding is they scuttled slash changed their organizational structure around AI and kind of recognize that it should have been their year, but it wasn't. But boy, you know, in classic Apple fashion, putting aside the prediction of 25, having in 2025 or 2026, think about how they're poised right now.
If they indeed can put Gemini 3 into their system and do it now, better late than never. And they've got the distribution, they've got the platform. it's an exciting prediction to sort of recycle it and say, 2026 will be the year of Apple AI. I think I'm gun-shy to say that because of what happened. But in terms of 2025, they didn't do very much at all in terms of the AI.
Andy Sack (08:47)
Yeah, we
missed this one. I think it was wishful, fanciful thinking. And I know they just cut the deal with Google and Gemini 3, but it may be wishful, fanciful thinking again this year. Yeah, can't count Apple out, but based on the last two to three years performance, it doesn't bode well for 26.
Adam Brotman (08:59)
Yeah.
Andy Sack (09:08)
The next prediction is around generative video and AI creators. And we said that 2025 would be a breakout year for generative video and AI powered influencers.
Adam Brotman (09:22)
It was a well, again, think the theme of all these things in review is that we kind of got it right and kind of got it wrong. the the it was a big year in generative AI video, like huge year like VO three, I believe was this year. ⁓ Sora to you know, we we've in our boot camps and in our consulting practice, we often show off the incredible stuff from like the call she add and
Andy Sack (09:36)
Yes.
Adam Brotman (09:48)
and the IM8 ad. so it was a breakthrough year in video for sure. Did it give rise to the AI video influencer in any sort of meaningful breakthrough way? No. it was a good thought on our prediction. it wasn't like in 20, it's like, my God, look at all these video AI influencers that are now prominent. And it didn't happen yet. But it was a big year for AI video.
Andy Sack (10:11)
Yeah,
think that's right. think big year for AI generative video. And I know AI influencers rose, but it was not nearly to the level that we were predicting it. I thought that AI influencers would take over. And that has not happened. There are AI influencers and AI bands making music, but it hasn't taken over just yet. I think the economics of that shift will continue to play out in 26.
The next prediction was regarding AI agents for everything. And we thought that 2025 would be the year of agentic AI and that many people would be managing employees and agents to get work done.
Adam Brotman (10:53)
You know, we got like the theme of this whole review, we got that a little bit right and a little bit wrong. Like, so it's interesting,
AI agents, the word agents, and when we made this prediction, we were at fault in ourselves. Like that word is like a very ambiguous, vague word. It gets used in lot of different ways.
The number one AI agent use or agentic behavior, in my opinion, are these reasoning models that came on at the very end of 2024 and have just dominated in 2025 in terms of the impact of these reasoning models on the outputs of these models. You've got thinking models or reasoning models, whatever you want to call them. They...
They feel agentic deep, deep research, as you pointed out, I think came out in February, you said, which I thought I was like, no, it had to be before that. And I was wrong. Like you were right. It was in February. So deep research is an incredible agent. It's a variation of a reasoning model that does a specific agentic task called research and rights and does analysis. But just like all these reasoning models now, like it was the year of the reasoning model. So if we would have said, ⁓ man,
I think ⁓ one and and reasoning like reasoning answers to chats is going to dominate the landscape. We would have been right. We were talking about real autonomous agents though that come along and do a bunch of like tasks for you and stuff and that did not materialize the year of the agents as most like including us were prognosticated did not materialize in a major way but this sort of precursor foundation level to them with reasoning models did.
in a way that we didn't predict, but it's still agentic. So we were sort of right, but we got it wrong, I think. What do you think?
Andy Sack (12:47)
I agree. I think that agent was definitely the buzzword of the year and the overstated promise of the year. So I think we mostly got that wrong. do think, and mainly like I had envisioned that both.
in the enterprise, there would be people working with agents and managing employees and agents, as well as that you'd be able to book a hotel and travel with agent software. And that just hasn't happened yet. So it was a great buzzword for the year, but I'm going to say that we got that one wrong.
Adam Brotman (13:22)
No, that's worth pointing out, though. Do we have any more? we kind of? We got more. I want to say one thing, though, on the agent thing that's worth highlighting. That we, where it was the year of the agent in AI was in coding, probably. I actually think if you talk to most software developers, they'd say, man, the difference from the end of 24 to the end of 25 in terms of like,
Andy Sack (13:27)
yeah, no, we've got a few more.
Yeah.
Adam Brotman (13:49)
the Claude code and Claude just released a new Claude Opus 4.5, like the capability of cursor with Claude or whatever your coding agent of choice. And they're agents. In that regard, we actually had ⁓ one of our colleagues tell us that 2025 was the year that you went from mostly writing code to mostly reviewing code.
And that's a line that like, think coding agents did. It was the year of coding agents, I think you could say.
Andy Sack (14:14)
Yeah.
Yeah. OK, but I think when we said it, we weren't thinking about coding agents. So I'm going to continue to give us that one we were incorrect on, even though your caveat is noted. I'll take this next one, AI First companies as the breakout story in 2025. And we predicted that.
Adam Brotman (14:25)
No, but we weren't... ⁓
Yeah, I think we were incorrect on that.
Yeah.
Andy Sack (14:44)
In 2025, you'd see the AI first application layer winners emerge and be easy to point to. And I'm going to say we're still early. I mean, it's funny that we definitely have that point of view individually and as a company at Forum 3. we're...
I think we're still early on the application layer. I 2025 was really the year of massive infrastructure build out. was all the talk was about Nvidia and the chip companies and energy. It was not about the application layer. So I'd say we're early on that and we got that one wrong.
said, we started the year with the interview of Moderna. If you haven't listened to that episode, I highly recommend it. It's a great episode. And we still firmly stand by the prediction. We were just early.
Adam Brotman (15:40)
We definitely got it wrong in the way that we phrased it. I think that you're rubbing off on us and you encourage us to have these bolder, more interesting predictions. And then they end up being sort of directionally interesting but wrong in their boldness. The one thing that, and I'm guilty of that too, that the one thing I'd say is, wasn't it this year Andy that there was like that
fate of AI First memos that came out. Maybe that was 2024, but I think it was early in 2025 that we saw like 10 CEOs come out. Remember that? And say, we're gonna be AI First. And they wrote the memo. that's not what our prediction was. But it's worth noting that there was, interestingly enough, using the term AI First, a whole bunch of companies that declared themselves to be AI First, which is different than
It happened and you can spot it and their stock went up and they out competed and all that stuff. That didn't happen, but the AI First memo from the CEO was a thing in 25, I believe.
Andy Sack (16:45)
Are there things that happened in 2025, like one or two things that you want to highlight outside of the predictions? That's the of the review of our predictions. What do you want to call out for the audience in terms of what happened in 2025 that was not included in the predictions?
Adam Brotman (17:08)
Well, we didn't predict. I'm trying to, I'm gonna give you a couple. I'll give you a couple. Well, let me give you a couple.
Andy Sack (17:13)
You want, you want me to, you want me to, you want me to tee you up a little bit? Yeah. So vibe, vibe
coding, vibe coding.
Adam Brotman (17:21)
Yeah, vibe coding was not on our prediction list. didn't, I don't think deep seek was the thing that was happening in December. And I believe December and January. So we weren't, it wasn't on our radar in November, December as much. And I, don't think we talked about deep seek and like the rise of like China's capabilities. we didn't, we definitely didn't predict the strength of Google.
and Gemini and sort of, if anything, I think people were sort of on the other side of that. We weren't predicting their demise. We were not as much on top of Google. think, yeah, you mentioned vibe coding. And I would say that I don't think we also
We weren't as clear at the time of the infrastructure layer build out and sort of the massive amounts of money that we're going to go into, you know, the Stargate or whatever it's called project from OpenAI and Oracle and like NVIDIA has continued just unbelievable run and like the amount of evaluation that was going to get propped up through AI infrastructure.
Andy Sack (18:41)
Yeah, two things that I'll-
Adam Brotman (18:42)
promises.
Andy Sack (18:44)
Two things I'll just chime in and add on that. One is at the beginning of the year, we definitely weren't talking about an AI bubble. And there's no question that 2025, as we end the year, the AI bubble is impossible to ignore on CNBC and in the press. And all the vendor deals and accounting and Michael Burry sort of pointing out that NVIDIA and this, like all of that is,
was not on the horizon. The other thing that was that I think we saw at the beginning of the year, but we didn't call out what's really just how competitive the market was is amongst the large language models. And, and that's really playing out. It's narrowed a little bit, I my guess is it will continue to narrow, but you've got a fierce competition going on. And we didn't met his billion dollar employment offer.
to get turned down by one employee. mean, that stuff was just crazy in 2025.
Adam Brotman (19:47)
Yeah, the yeah, if you I mean, if you just sit back and think about 2025, yeah. And you think, OK. Apple did, I mean, you can almost do it by by company or by topic, but like Apple did not like you think about 2025 was like the dog that did not bark was Apple. That was interesting. Metta, who's always been on top of stuff like it's interesting, you know.
We would have never predicted like Yon LeCun would basically be leaving. They would buy scale AI and reorganize their entire world. like, we're not talking about llama anymore. Like llama was a major model at the time last year. like, so like Apple and Meta kind of had a year of wilderness. know, open AI for all the fact that they're ending the year with a little bit of, you know, there's kind of a negative vibe on some things with open AI. Open AI had an amazing year.
Andy Sack (20:23)
That's right. That's right.
Adam Brotman (20:39)
if you think about it overall, it's just not, it's just not ending. Yeah. I mean, just like their user growth and the things that the deals they were involved in, like, you know, it's just that we're right now we're recording this episode, right? When Google just spiked the ball in the end zone on a bunch of stuff. And, but the truth is like opening out a great year, Google had an amazing year. Elon Musk had an amazing year.
Andy Sack (20:41)
Amazing. Unprecedented.
What, Adam, what's your,
what's your take on Microsoft? How was their year in AI?
Adam Brotman (21:08)
Great question.
I, you know, it's sort of, it's hard for me to answer because it's, and I think it's reflective of how I, how I think their year was. first of all, let's, let's just look at the numbers. They're, they're Azure cloud to support AI solutions for their clients, I'm guessing, which is what I'm understanding is a lot of the demand is to use cloud computing from Microsoft using Azure for AI applications, which is like their bread and butter in the enterprise.
was incredible. They had big numbers, their stock is up for good reason. They've managed that business incredibly well. I think I tend to get caught up. We tend to get caught up in the fact that Copilot itself doesn't have the best reputation by enterprise users. And it does perfectly fine. I'm not knocking on Copilot. But if you ask many enterprise users that we talk to every day, they'd be like, oh, man.
I like chat TBT better than copilot for like general analysis and chat. And I think we hear that a lot, I, my way, that's my way of answering that. I think Microsoft had a good year, but I think it's, it's in the Microsoft. You got to get out of the mindset of the sort of flashy either consumer or like the latest model. They don't have their own model. That's really well known if at all, right. That we know about they they're mostly using open AI under the hood. they're
building their cloud computing, copilot for all the fact that I just mentioned it does what it does. I don't think, I bet you if you looked at the numbers are probably outstanding. So that's my way of saying like at the end of the day, think Microsoft quietly had a very good year, but it wasn't as flashy as some of the other stuff in the drama. How would you answer that?
Andy Sack (22:44)
Yeah.
I don't have much to add other than to say that I know we hear from clients that their productization of AI, which I think is really hard, which is really OpenAI's advantage, is that they're a new company, they're operating at AI speeds, and they're productizing, frankly, better than anybody in the market. Microsoft has always struggled with that until somebody shows the right way to productize and they can copy it.
the integration, even though they have access to the open AI technology, the productization we've heard from clients is just not great. And so yes, the cloud side of their business has continued to explode and that has really put them in position to compete with AWS. So I'd say it's, it's been a good year, not great. And I don't feel like they executed nearly as well as open AI, Google or
or Claude or a DARPAG I ⁓
Adam Brotman (23:51)
Yeah, and we should
and we should just make that one comment like at the end of 2025, like, and thropic they were they were it was the word out as they have a b2b business and open AI, probably based on the back of cloud code. Their revenues are up there actually, did you hear that they're on track for like profitability in like two years? Like, I got a hand at tantropic they they might be the dark horse quiet, you know, you know, right up there.
Andy Sack (24:11)
Yeah.
They are. think they are. ⁓
We have just a moment left. So I'm going to let's give final comments on the year and then I'll wrap up. So you first. Anything you want to add?
Adam Brotman (24:31)
No, I think we covered it.
Andy Sack (24:33)
I think we covered it too. mean, it was a unprecedented year. think the thing that we didn't, that I would say that the labor markets, you know, employment continues to be strong. We did not really see the impact on employment in 25 that I think we would have forecasted at the beginning of the year. We did see a slowing of hiring.
And then the other thing is that the regulatory picture as it relates to AI continues to be murky with the state's rights versus the feds and really the competition with China. So those are two areas that are worth noting at the end of this episode and the end of 25.
Adam Brotman (25:14)
Yeah, actually, the only thing I wanted to say is that.
I think the fact that the way that the year in review went compared to our predictions, I think is a sign of how 2026 is going to go. think quietly we're going to be amazed by the progress. And yet I think making bold predictions of rapid societal changes are unlikely.
⁓ I you know, I at some point maybe they'll just flip and it'll happen But I don't I think that was a lesson I learned in this episode going through this with you is like I think we'll continue to see progress and yet I don't think it's gonna be like 26 was the year of the whatever. I I just think these things are You know continuing to sort of boil us like a frog a little bit
Andy Sack (26:02)
Yeah, yes, we're being boiled like a frog. But I think when you all hear our predictions for 26, I think we're on a curve. So there's not going to be cliffs. But I think you are going to see those two issues, both regulations and labor, significantly more in 26 than we did see in 25.
And these episodes in the year and reviewing them is just a really fun exercise. I hope all of you that are listening enjoyed it as much as Adam and I have With that, thank you all for listening to AI First with Adam and Andy. For more resources on how to become AI First, you can go to our website, forum3.com, download case studies, research briefings, executive summaries, and join our email list. Also, we want to invite you
to connect with us on our AI First community, a curated hub and network for leaders turning AI hype into action. We truly believe you can't over invest in your AI learning. Onward.