0:03 The stock market is just melting down. The AI bubble just popped. The big tech
0:07 stock crash is just beginning. And here's what caused the whole thing. The
0:12 people on LinkedIn found out about Claude Code. That's a joke, but it's not
0:16 too far from the truth. Let's unpack exactly what happened. So, starting in
0:20 about November of 2025, just a few months ago, a lot of people in the AI
0:24 tech space, the AI sphere, so to speak, most of us were kind of losing our minds
0:28 over cloud code. We saw the massive possibilities there, but most people
0:31 that weren't developers just couldn't get their hands on it. It had a steep
0:36 learning curve. Then Anthropic releases co-work and co-work for normies, if you will. It was the
0:43 evolution of claude code, but for everybody and at the time it didn't make
0:47 that big of a splash, but that was just the come before the storm because this
0:52 week Anthropic released a set of tools or kind of the starter plug-in set for
0:58 co-work. So, Anthropic drops 11 starter plugins on GitHub that people can use to
1:03 do their work in in co-work. So, that happened January 30th. The thing that
1:07 was kind of like the the catalyst, kind of the the the lynch pin that set the
1:12 whole thing going was specifically the legal plug-in. So, it included a legal
1:16 plug-in to do various legal work. You can use commands like review contract.
1:20 So, it would take a contract and review it clause by clause against your
1:24 playbook, kind of the context of what your firm does. and you take any given
1:28 contract and the claude co-work would go clause by clause and it would return
1:32 green, yellow, red flags depending on the severity of conflicts etc. Another
1:37 slasht triage NDA. So, if you're in a position where you get tons of NDAs that
1:40 you need to sign because you're dealing with some secretive technology or
1:43 whatever, these basically triage them into like, yeah, standard boiler plate,
1:48 go ahead and sign it into, hey, this is mostly okay, but here are a few sort of
1:51 concerning points that you have to address versus, uh, this needs like a
1:55 council review. This needs your your lawyer to take a look at it because
1:58 there's a whole bunch of can of worms in there. There's one that did a vendor
2:02 check, checking vendor agreement status, briefs, responses, etc., etc., etc. And
2:08 that hit a lot of companies kind of hard. One of the biggest examples is
2:11 Thompson Reuters. It's a company that does about six billion a year providing
2:16 various legal services, advice, research, etc. The entire business model
2:21 is based on lawyers paying expensive subscriptions in order to get access to
2:26 databases, various other research things that people in that profession need. So
2:32 Thompson Reuters takes a 20% hit in like one day. it stock price drops 20%.
2:38 Because this one plugin on GitHub for Claude Co-work basically, you don't want
2:42 to say it completely replaces them, but man does it take a big big chunk out of
2:47 what they do. And the cost for it goes from I I assume very expensive. That's
2:50 what they charge the lawyers probably some high fee to, you know, you're just
2:55 dealing with claude that has a plugin. So it's either free if you're in on one
2:59 of the plans or if you're paying API cost and it it's still going to be very
3:02 very cheap. But this isn't just about legal review or whatever illegal
3:06 companies. Really, what I think it made people realize is that vibe coding is no
3:11 longer this funny term that people throw around. It's no longer AI slop. It's
3:16 capable of creating incredible software that is free. Free in the sense that I
3:20 can ask cloud cod to generate some software for me. If you seen my video on
3:24 cloudbot aka open claw, I mean you saw I was building what could be considered I
3:28 don't know if it's enterprisegrade software. It's what would have been a a
3:32 good SAS company, but instead I'm able to just tell Cloudbot to create for me,
3:37 and it does. It's customtailored for my needs. It doesn't have a high price tag.
3:41 It it's made in less than 5 minutes. And if you've been following this channel,
3:44 we've been talking about it for at least a year plus. I know this because just a
3:48 few days ago, it was the one-year anniversary of Vibe Coding. When Vibe
3:52 Coding came into existence, at least the word, right, when Andre Karpathy said,
3:56 you know, let's call this thing Vibe Coding. That was a year and and a couple
3:59 days ago. So the fact that it happened uh you know days after the anniversary
4:03 is definitely ironic. And we're also seeing a lot of other tech stocks
4:07 specifically ones that bets on software either produce software or invest in
4:11 companies that rely on software production of software software as a
4:14 service etc. So people are calling this the SAS apocalypse SAS apocalypse. The
4:19 fears that what used to be billion dollar companies with competitive emotes
4:23 and now they're just poof they're evaporating. Of course, we've been
4:26 talking about it for again a year plus. A lot of these companies that make a lot
4:30 of money selling software. As AI coding gets better and better, that model
4:34 becomes unsustainable. So, they either have to adapt and somehow take AI into
4:39 account or yeah, maybe just cease to exist. Now, of course, this is very
4:43 likely probably I expect it's a overreaction to what's happening. If
4:47 you've been doing a lot of VIP coding, building the software, you know, there's
4:51 still tons and tons and tons of issues. It's excellent in certain use cases, but
4:55 there's still problems. One of the big problems is it kind of creates this
4:59 ticking a time bomb for technical debt. Right? If I'm sitting there just vibe
5:03 coding everything, the entire structures of of the business, the more and more I
5:07 build it, the less and less I sort of don't know about it, how were certain
5:11 features implemented, what are the security vulnerabilities that are that
5:15 are possible? The code that's created will often lack security scalability. It
5:20 might not have the best documentation. Most people don't also ask their AI
5:23 chatbot to provide the documentation to go along with everything. So there's
5:27 this kind of massive and growing cost for the future, the technical debt. This
5:32 not to mention the liability of various security weaknesses that that these that
5:36 this code can have. So it's not like AI coding by coding is just going to erase
5:41 every SAS company in existence. But the flip side of that is that a lot of
5:45 startups are getting a lot more shots on goal. Meaning they can quickly create
5:49 their sort of minimal viable product, quickly deploy it, put it out there.
5:53 They can do it cheaply without too much technical talent. The cost to create a
5:57 bare minimum product went from hundreds of thousands of dollars to basically
6:02 zero. And when you're just trying to get that version 1.0 0 and you're just
6:05 trying to get off the ground, get the first few users on board. All the
6:10 problems with vibe coding, they're not the biggest concerns. You're just trying
6:14 to gain some momentum, see if there's a product market fit. If the thing takes
6:18 off, well, then you can figure out how to try to, you know, build everything
6:21 correctly, reverse engineer it. Yes, that that might cost more money in the
6:25 future, but you also get to quickly iterate, test, see what works. And then
6:29 when you catch on to something that works, then you can pour the the talent,
6:33 the money, the technical resources into it. And it's not just for startups. It's
6:37 for anybody. Anybody with, you know, cloud code or or open codex can start
6:42 building some pretty impressive stuff with you need just a little bit of
6:46 technical knowledge, but it's not a lot. You can be barely technical and and
6:50 still manage to create some pretty amazing stuff. But the point for these
6:53 big software as a service companies is that while the difficult things that
6:58 they do aren't going away, a lot of them do make a lot of money on very very
7:03 simple software. And now that any kid can create their own versions of that
7:07 simple software for their own needs, that's going to impact the companies.
7:09 It's not going to evaporate them. It's not going to kill them. Not in the short
7:14 term, but a lot of the high valuations they were kind of projecting long-term
7:18 growth, profitability, etc. This is going to grind them down a little bit.
7:22 The other kind of big important related thing that was happening today is
7:27 somebody leaked the news that Anthropic will be releasing their new models. So
7:31 that's Sonnet 5, which is supposed to be better, faster, cheaper, smarter than
7:35 Opus 4.5. That's kind of the next generation of models, but the the
7:39 smaller version of it. And if that's the case, that would be kind of a big deal
7:42 for everybody that's using these AI agents for our ability to build
7:45 software. Not only do we see an improvement, but also the cost drops
7:49 massively, the speed increases. And not only that, but apparently we're also
7:54 going to be getting Opus 4.6, the next iteration of their big model. And we
7:58 were all expecting it today. And there were some things happening behind the
8:00 scenes that sort of suggest that this is the case. We're beginning to see
8:03 references to those models, you know, kind of a behind the scenes and
8:08 anthropics websites and API calls, etc. But something happened where we're not
8:12 sure what either there was a last minute decision to not release it. that it
8:15 wasn't ready for the release or maybe whoever leaked the information knew when
8:19 the infrastructure was going to start changing but that wasn't the actual
8:22 release date. Often times they build the functionality into their infrastructure
8:25 before the release. So you can test it, make sure everything's working properly,
8:28 etc. and then days later, weeks later gets rolled out. The point is 5 and
8:33 probably open 4.6, they're probably coming and coming pretty soon. So it'll
8:38 be interesting to see how these software as a service companies, how they
8:42 survive. I don't expect a full meltdown. And I'm I'm sure this is a bit of an
8:47 overreaction, but over the long hall, I would expect the valuations of these
8:50 companies to start coming down a little bit. As more and more people are
8:54 empowered to create their own software, you know how it just conjured into
8:58 existence for their specific use case. I don't see how that doesn't have an
9:02 effect on the market. I've replaced several SAS subscriptions by building my
9:06 own software using, you know, clot code, clawbot, open claw, and openi codeex.
9:11 Mostly I'm using the anthropic models for coding right now. And I know that
9:14 there's right now a lot of people like that. People that are watching, you have
9:17 probably experimented with that. You you might be doing a lot with it. We're like
9:21 the the weirdos, the freaks. We're a little bit in front of this wave. But as
9:25 more and more people realize what these things are capable of, more and more of
9:27 them will choose to create their own software. If you're able to just talk to
9:31 a chatbot with, you know, natural language and it creates the software
9:35 that you need. Why would you pay someone else to create it for you or use
9:39 whatever preset pre-programmed software they have? You might need that for
9:43 enterprise cases or specific cases where security is of high concern, but there's
9:47 a million use cases where something super simple is all that you need.
9:51 Anyways, it's very kind of exciting and interesting to see as like all the stuff
9:55 that we've been talking about for I mean a few years now. It's slowly sort of
10:00 percolating into kind of like the the everyday people awareness like oh these
10:04 things can code and yes they're getting to the point where they can code better
10:07 than a lot of people and yeah they can probably replace some of the work that
10:11 we're doing cuz remember not that long ago a lot of people would say that we're
10:15 crazy because we believe that this stuff is going to happen. Today I checked my
10:19 online analytics for some of my websites and realized that Chad Gypt is now one
10:24 of the top referers to my website more than Bing or YouTube or or many other
10:30 search engines. Chad GPT is now sending quite a lot of traffic to my website. So
10:34 fun times ahead. Let me know what you think about this. If you are investing
10:38 in publicly traded companies, who do you think are the winners here? Who's going
10:41 to be the benefactor of all that is happening right now? Let me know in the
10:44 comments if you made this far. Thank you so much for watching. and I'll see you
