Latest / In the Loop AI / Anthropic Just Leaked Their Entire Claude Roadmap by Accident (What's Coming Next)

Transcript

Anthropic Just Leaked Their Entire Claude Roadmap by Accident (What's Coming Next) · In the Loop AI

0:00
Josh: This morning, a security researcher posted a single link on X,
0:03
Josh: and within hours, it had 3 million views and had millions of copies backed up all across GitHub.
0:08
Josh: By the afternoon, when we're recording this episode, Anthropica is scrambling
0:11
Josh: to delete old versions of their NPM package, but it was too late.
0:15
Josh: What leaked was the entire source code of CloudCode.
0:19
Josh: Every single line, 512 lines of TypeScript, 1,900 files, every tool,
0:25
Josh: every permission system, every internal codename was leaked,
0:28
Josh: all because someone forgot to include a single debugging file from a public package.
0:33
Josh: And that story alone would be a major story. But what makes this even more crazy
0:36
Josh: is what people found buried inside the code.
0:39
Josh: And now we have information about every feature that's coming down the pipeline,
0:41
Josh: as well as all of the secrets that Anthropic and Claw team didn't necessarily
0:45
Josh: want us to know. This is a really big leak. I can't believe this happened.
0:48
Ejaaz: I mean, big leak is one way to describe it. Absolutely
0:51
Ejaaz: terrible for the Anthropic security team is another
0:55
Ejaaz: one brutal everything this is the second leak um
0:58
Ejaaz: that anthropic has made in the last five days so
1:01
Ejaaz: they're shipping our new product every single day but they also seem to
1:04
Ejaaz: be leaking their entire roadmap we now know what the next 44 product releases
1:09
Ejaaz: are going to be over the next couple of months or rather than a couple of weeks
1:12
Ejaaz: for anthropic right now for code code specifically as you mentioned half a million
1:17
Ejaaz: lines of code 19 000 files and a bunch of different feature releases,
1:22
Ejaaz: which, by the way, have already been built. So they just need to click the launch button.
1:26
Ejaaz: We have all the details and we're going to get into it. But before we do that,
1:29
Ejaaz: We need to kind of describe how this happened because leak is one way to describe
1:34
Ejaaz: this, but it wasn't an internal employee at Anthropic leaking these files or this source code.
1:39
Ejaaz: This was publicly available. Let me repeat that.
1:42
Ejaaz: This code was publicly available in the latest update of clawed code.
1:46
Ejaaz: Someone within Anthropic had mistakenly left a file, a .map.js file in the system
1:53
Ejaaz: that was publicly accessible.
1:55
Ejaaz: Someone found it. And now that original post that exposed this source code has
1:59
Ejaaz: been seen by over 10 million people, and it's only been three hours since it
2:03
Ejaaz: got posted as we're recording this, and it's been forked over 5,000 times.
2:08
Ejaaz: So this is basically Claude Code's entire blueprint, entire architecture,
2:13
Ejaaz: the way that its memory is set up, the way that the model works,
2:16
Ejaaz: released for anyone and everyone to use.
2:18
Ejaaz: And a bunch of people have already been using it. People have plugged in different
2:21
Ejaaz: models, have created their own versions of Claude Code.
2:23
Ejaaz: It is just insane. Josh, even you forked it this morning, right?
2:26
Josh: It's amazing, yeah. And so just to clarify, CloudCode is Anthropics command line tool.
2:30
Josh: This isn't the full CloudDesk type application, but it's a tool that lets developers
2:33
Josh: talk directly to CloudCode in their terminal.
2:36
Josh: And it's very powerful software. So what happened, like you mentioned,
2:39
Josh: and this matches the pattern of the previous leak that we covered in yesterday's episode about the new
2:46
Josh: issued the code themselves it was just available publicly and
2:49
Josh: the problem is because when they publish this code there's an
2:52
Josh: npm package containing this like dot map file and it's a source file that references
2:56
Josh: the complete source code and that source was directly downloadable as a zip
3:01
Josh: file from anthropic's own cloud storage bucket you just went to anthropic you
3:05
Josh: asked them hello sir can i please have the map file that tells me where all
3:09
Josh: of these references go to and they delivered it to you.
3:11
Josh: And the irony here is that Anthropic built an entire subsystem called Undercover
3:16
Josh: Mode, specifically designed to prevent internal information from leaking.
3:19
Josh: And it does things like strip the model code names and the project names,
3:22
Josh: and then went ahead and leaked everything through a build configuration oversight.
3:26
Josh: And it's really got to be code red. If you're waking up at Anthropic right now
3:29
Josh: as a developer, this must be a really brutal morning for you.
3:33
Ejaaz: The funniest part is the Undercover Mode that you just mentioned was literally
3:37
Ejaaz: meant to obscure or all of this.
3:39
Ejaaz: And the fact that they exposed it publicly means that whoever got access to
3:43
Ejaaz: it could just reverse engineer the entire thing.
3:45
Ejaaz: So let's say you gave Anthropics new model a code name, you could reverse engineer
3:49
Ejaaz: the file to find the original name of the model and how it works.
3:53
Ejaaz: It's just been the craziest mess up in Anthropics so far.
3:56
Josh: So now let's get into the good stuff. This is what's coming down the pipeline.
3:59
Josh: If you are a user of Cloud Code or Anthropics products in general,
4:02
Josh: we have the totally unreleased roadmap now in plain text available to walk through.
4:07
Josh: And I think that's what we're going to do right now.
4:09
Josh: Ijaz, you have this nice little artifact generated by Claude Code itself to
4:13
Josh: walk us through all of these new features that are coming to one of our favorite
4:16
Josh: products that we use every day.
4:17
Josh: So please, let's hear the leaks. Let the leaks flow. Let's see.
4:21
Ejaaz: Thank you, Claude Code, for creating your own demise and a beautifully visual
4:25
Ejaaz: artifact for this episode. Thank you very much.
4:27
Ejaaz: So at the start of this, or at the top of this page, it says there were 44 product
4:31
Ejaaz: releases that people had never heard of before.
4:33
Ejaaz: So everything you're about to hear right now is new. Okay.
4:36
Ejaaz: There were 20 specific product releases that caught people's attentions,
4:40
Ejaaz: and we're going to go over the top ones for you right now.
4:43
Ejaaz: So the first product release is called Kairos, which is basically an always-on autonomous Claude.
4:50
Ejaaz: What that means is when you use Claude code, you typically have to monitor it,
4:54
Ejaaz: come back, check the code, make sure it's doing the right job, test the code, etc.
4:58
Ejaaz: This new update will basically allow Claude to autonomously run on its own.
5:02
Ejaaz: It can check its own tasks.
5:04
Ejaaz: It could create new tasks for itself and work towards a goal.
5:07
Ejaaz: So you could leave it unattended for hours and hours at a time. It's pretty awesome.
5:11
Josh: What I found cool about this also is Kairos will do nightly dreaming,
5:15
Josh: So a forked sub-agent will run four phases. It'll orient, gather,
5:19
Josh: consolidate, and then prune, and then distills these daily logs into these structured topic files.
5:24
Josh: And then overnight, it will bake them into the memory and actually learn the
5:27
Josh: same way that humans do, where overnight it will dream and then lock this into
5:30
Josh: the memory and grow and get better every single day. So Kairos is very cool.
5:33
Ejaaz: But this next one is my favorite. This is so cool. This is so cool.
5:37
Ejaaz: So it's codenamed Buddy, and it is basically a virtual pet AI companion that
5:44
Ejaaz: lives on your CLI, on your command line interface.
5:47
Ejaaz: It's meant to, and this is me guessing here, act like a personal AI agent assistant
5:52
Ejaaz: that can assist you on all things coding related, but also once you publish
5:56
Ejaaz: the code, helps you edit the app, review the app that you created,
6:00
Ejaaz: walk through it, find bugs.
6:01
Ejaaz: Basically, it's a personal assistant that lives on your computer and off your
6:04
Ejaaz: computer when you're publishing artifacts or whatever that might be.
6:07
Ejaaz: This reminded me of a game, Josh, and it says it on the screen here,
6:11
Ejaaz: Tamagotchi, which we, I don't know for the age of the audience or listeners
6:15
Ejaaz: here, but we used to have these like cool devices that you can kind of like
6:17
Ejaaz: hold in your pocket or in your key chain and you had to keep the virtual pet alive.
6:22
Ejaaz: This reminds me of that and Microsoft Clippy. Do you remember Microsoft Clippy,
6:26
Josh: Josh? Very well. I love having companions. And we have some additional information
6:30
Josh: about this buddy system in that there's 18 species of buddies and a lot of them are animals.
6:36
Josh: We have ducks, gooses, blobs, cats, dragons, octopuses.
6:40
Ejaaz: Capybaras.
6:41
Josh: Is there a capybara? There is a capybara. Interesting.
6:45
Josh: And actually what we're seeing on screen now is someone took this information
6:48
Josh: and kind of rendered what he presumed it would look like. So you choose your species.
6:52
Josh: Each species of animal has a rarity tier. There's common, uncommon,
6:56
Josh: rare, epic, legendary, and then there's shinies even.
6:59
Josh: So it's like this whole tiered game that's built on top of it.
7:02
Josh: And then there's Statistics like debugging, patience, chaos, wisdom, snark.
7:05
Josh: And what you're seeing on screen is this person's kind of choosing his character.
7:09
Josh: He's choosing the traits that it has.
7:12
Josh: I assume there's some sort of rarity baked into this. And it's going to be this
7:15
Josh: fun gamified version of a Tamagotchi built into quad code, which seems really
7:21
Josh: interesting and novel. And I don't know, it just seems fun.
7:23
Josh: Did you say that this was first releasing tomorrow.
7:26
Ejaaz: Josh, April 1st? Do you think this is like a joke?
7:30
Josh: They're teasing this on April 1st for release in May.
7:33
Josh: So if that's true, by the time you're hearing this episode, within an hour or
7:37
Josh: so, they should be teasing this.
7:39
Josh: If the leaks are true, if they don't change their mind, And then if that's true,
7:43
Josh: then the odds are that this will release in May is probably correct,
7:46
Josh: because that's what's said in the code.
7:47
Josh: Now, like you mentioned, tomorrow is April Fool's Day, or I guess when you're
7:50
Josh: listening to this, happy April Fool's Day.
7:52
Josh: And there is a chance that this isn't true. But I based based on the rest of
7:55
Josh: the leaks, it seems like this was very much not intentional.
7:57
Ejaaz: Okay, but there are three more features that I want to get through as well.
8:01
Ejaaz: One of these is called coordinator mode, which basically describes a multi-agent
8:06
Ejaaz: program that allows you to control a swarm of AI agents.
8:11
Ejaaz: So right now, it's typical if you're a software engineer to spin up not just
8:15
Ejaaz: one instance of code code, but multiple.
8:17
Ejaaz: That's normal. People are already doing this. But an issue starts to arise when
8:20
Ejaaz: there are multiple of these agents. We're talking like 50 plus,
8:24
Ejaaz: 100 plus that are doing all different types of work and need to kind of work
8:27
Ejaaz: together to figure problems out together. It becomes really hard to coordinate.
8:31
Ejaaz: This coordinator mode is basically Anthropix feature to help you manage all of these.
8:36
Ejaaz: Think of it as like an operator board or a control system that you can kind
8:39
Ejaaz: of like manage it, similar to like a strategy computer game.
8:41
Ejaaz: It's funny, there's a lot of like computer game analogies in the features that
8:44
Ejaaz: they're releasing. This is basically that.
8:46
Josh: There's also one that I really enjoyed, which is the Ultra Plan feature. Oh yeah.
8:49
Josh: And it basically solves the problem of Claude running out of context by giving
8:54
Josh: it a 30-minute sandbox in the cloud to think deeply before presenting a plan.
8:57
Josh: So when you're working on these complicated things with Claude code,
9:00
Josh: it often refers to plan mode.
9:02
Josh: But plan mode sometimes runs out of context. It doesn't have all the information.
9:05
Josh: This offloads all of that in a 30-minute window.
9:08
Josh: To a giant server that can handle all the context and actively improve the planning
9:13
Josh: of the project that you're building.
9:14
Josh: So when you go and set it free to go build these things, it has a much better
9:17
Josh: idea of exactly what you want.
9:19
Josh: And I think plan mode, if you're building anything serious, is a really powerful
9:22
Josh: thing. And adding ultra plan on top is something that I will be using very much
9:26
Josh: so for the larger projects.
9:27
Ejaaz: That's such a good point, because right now, they keep on promoting that Claude
9:31
Ejaaz: has or Claude has like a 1 million context window, but it becomes super crappy
9:36
Ejaaz: after 200,000 characters, right?
9:38
Ejaaz: So like the performance quality goes down. So this is hopefully something that fixes that.
9:42
Ejaaz: So I'm excited to see that in the pipeline. But there's one more thing that
9:45
Ejaaz: I want us to talk about, which is called or referred to as the custom agent
9:49
Ejaaz: creator, code name wizard.
9:51
Ejaaz: So typically when you set up Claude Code and you use Claude Code,
9:55
Ejaaz: you're using the system prompt that Anthropic gave to you. It is like predefined.
10:00
Ejaaz: It is already written out. So you can't kind of adjust the personality of the
10:04
Ejaaz: Claude Code agent or anything like that.
10:06
Ejaaz: This new builder gives you that opportunity. You can form and create your own
10:10
Ejaaz: agents with their own personality, own memory types,
10:13
Ejaaz: different kinds of tools that you can give them access to, locations,
10:16
Ejaaz: or maybe they live on your desktop, or maybe they live in the cloud,
10:18
Ejaaz: or maybe they live somewhere else locally on a hardware device.
10:21
Ejaaz: You can control and manage all of these. Now, with the earlier product that
10:24
Ejaaz: I mentioned, which is the multi-swarm coordinator, you can start to see how
10:27
Ejaaz: these different pieces of the puzzle fit together to create some kind of gamified
10:31
Ejaaz: experience for end-to-end software engineering.
10:33
Ejaaz: It's just really cool to see all of this. But the craziest part about all of
10:36
Ejaaz: this, Josh, is all of these products and features are already built.
10:41
Ejaaz: They're built, they're just unreleased yet. So I'm starting to see why Anthropic
10:45
Ejaaz: or how Anthropic has been able to release a product every single day.
10:49
Josh: But we don't have that code. We can't actually create these buddies.
10:52
Josh: We can't actually use Superplan yet. We don't have everything.
10:55
Josh: So what was leaked today, it's probably important to distinguish what we have
10:57
Josh: versus what we don't. This is a huge leak, but it's not everything.
11:00
Josh: So if I were to download a copy on my computer, I would get the harness,
11:04
Josh: right? And Ijaz, you were describing it to me earlier as the car body.
11:07
Josh: We're not actually getting the brain. We're not getting the clawed model weights.
11:10
Josh: We don't have this brilliant intelligent model now that we could run locally,
11:12
Josh: but we do have the software that kind of acts as a harness for it.
11:15
Ejaaz: Is that For all of those people who are getting excited about getting access
11:19
Ejaaz: to the blueprint for Claude's AI model itself, this is not that.
11:25
Ejaaz: Think of the engine of a car being the actual model and the intelligence of the AI itself.
11:31
Ejaaz: And then think of the code that got released or leaked today as being the car
11:35
Ejaaz: chassis, the actual car body.
11:37
Ejaaz: So what's cool about this is, whereas you may not have access to Claude,
11:42
Ejaaz: the model itself, the code from that model, you can plug in an open source model.
11:46
Ejaaz: And people are already starting to do that. I'm seeing instances online where
11:48
Ejaaz: people have plugged in DeepSeek, they've plugged in Quen, and created their
11:52
Ejaaz: own version of Claude Code, the CLI interface and whatever that looks like.
11:55
Ejaaz: So this is really critical infrastructure and software.
11:58
Ejaaz: I cannot believe the Anthropoc team released this. It is just, it's so nutty.
12:04
Ejaaz: It's so bad. This is like, this is an IP issue right here.
12:07
Ejaaz: Like their equity, their $350 billion, actually rumored $450 billion private valuation.
12:13
Ejaaz: A lot of it is based off of claw code which has risen to extreme popularity
12:17
Ejaaz: over the last six months so it's just insane that this has actually happened
12:21
Ejaaz: there's more um product features are one thing 20 releases ready to go but we
12:26
Ejaaz: also got confirmation about the latest clawed models that are about to be released
12:31
Josh: Yes this is very cool for those who haven't seen our episode that we
12:33
Josh: just published yesterday it is all about the previous leak that happened with
12:37
Josh: claw which is called mythos and capybara the new internal model names and now
12:40
Josh: we have actual verification from the source code of anthropic that they are
12:44
Josh: here so what we're seeing on screen now is kind of like a system prompt for
12:48
Josh: this thing called undercover mode And now Undercover Mode is meant for Anthropic employees only.
12:52
Josh: When they use Cloud Code to publish on public and open source repos,
12:57
Josh: they use Undercover Mode to kind of strip away all of the classifying characters
13:02
Josh: that would possibly leak information out to the public.
13:05
Josh: So in this system prompt, it says,
13:07
Josh: never include commit messages or PR descriptions of internal code names.
13:11
Josh: For example, animals like Capybara or announce any unreleased model version
13:16
Josh: numbers like Opus 4.7 or Sonnet 4.8.
13:19
Josh: As I was reading this, I found one that I found particularly interesting at
13:22
Josh: the bottom of this under bad, where it says, bad, never write these.
13:27
Josh: Fix bug found while testing with Claude Capybara. And I was like,
13:31
Josh: huh, that's interesting.
13:32
Josh: Clearly they are using Capybara internally. And I have to ask,
13:35
Josh: is this the reason why they've been shipping product features so quickly?
13:38
Josh: Are they using this God tier model that they have internally that they've been
13:42
Josh: teasing that costs a tremendous amount of dollars per token?
13:45
Josh: And they're using that to actually just
13:47
Josh: build the code, review the code, and then publish it faster than everyone else.
13:51
Josh: It seems like that's possibly the case.
13:52
Ejaaz: I mean, in the words of Boris Cheney, the founder of Claude Code,
13:55
Ejaaz: he said a couple weeks ago, can confirm Claude Code is 100% written by Claude Code.
14:02
Ejaaz: So we know that the AIs are building the AIs.
14:06
Ejaaz: I think OpenAI is doing the similar thing with Codex. And that is the reason
14:10
Ejaaz: why these teams have been able to ship so quickly.
14:12
Ejaaz: Now, I wish I had a tinfoil hat nearby because I of a conspiracy mode,
14:16
Ejaaz: Josh, which is these AI models might be leaking themselves and it may not be
14:20
Ejaaz: the Anthropic engineers.
14:22
Ejaaz: I know that sounds insane, but I don't think it's unlikely. I'm going to put
14:27
Ejaaz: it at like maybe a 5% to 10% chance.
14:29
Ejaaz: But the point is, there are a bunch of new models being released by Anthropic coming up soon.
14:34
Ejaaz: We mentioned Capybara. We mentioned Mythos, which is meant to be these big,
14:38
Ejaaz: huge models trained on 5% to 10% trillion parameters, which is like a 3x increase
14:44
Ejaaz: in the size that we already are seeing and using with the models today.
14:47
Ejaaz: It's going to be an absolute beast of a model.
14:50
Ejaaz: It parinates a cybersecurity risk, which is incredibly ironic because all of
14:53
Ejaaz: that droplet stuff is getting leaked right now.
14:55
Ejaaz: But also Claude Opus 4.7 and Sonnet 4.8. So we're going to get version upgrades
15:00
Ejaaz: of the existing models that we're having already.
15:02
Ejaaz: So my one question is, when are these models going to get released?
15:06
Ejaaz: Because I need to get my hands on them.
15:08
Ejaaz: Number two, will it cause my entire laptop to get hacked? I don't know.
15:12
Ejaaz: So there's like a reputation risk going on right now as well as I want to use the actual thing.
15:16
Josh: Well, you also mentioned the security part of this, and I think it's worth noting
15:20
Josh: that there has been an increased cadence in security issues recently and leaks and exploits and hacks.
15:27
Josh: And I know they happen all the time, but I can't, like, there is some sort of
15:31
Josh: correlation happening here between models getting smarter and exploits.
15:34
Josh: I mean, yeah, we have this post on screen here, which summarizes it in a great way.
15:37
Josh: It says, this week in security, there is, what is that, six different exploits
15:41
Josh: that happened and pretty serious ones too.
15:44
Josh: Axios, which is a npm supply chain hack
15:47
Josh: that affects like many many millions of projects and applications
15:51
Josh: and if you've ever provide code or anything chances are you use that dependency
15:54
Josh: um openai codex had a command injection via github branch there's a terabyte
16:00
Josh: data leak from mercore and this doesn't even include the leak from today which
16:03
Josh: is cloud code so there's this increasing cadence of leaks and exploits and you
16:10
Josh: gotta ask the question is like if anthropic internally is using these tools,
16:15
Josh: Who else has access to tools this powerful? What can they be used for?
16:19
Josh: Are they actually responsible for any of this? Or is this just a random correlation
16:24
Josh: that's happening? I don't know.
16:27
Ejaaz: I think my main concern is that malicious scenario that you described where
16:32
Ejaaz: people are accessing this tool but using it for bad purposes is already happening.
16:37
Ejaaz: It's coming in the form of prompt injections.
16:39
Ejaaz: Like, look, there are six hacks that happened this week alone,
16:42
Ejaaz: and it's only been like two to three days.
16:45
Ejaaz: I wonder if that increased cadence is based off of people being able to get
16:50
Ejaaz: access to intelligent AI models like this and finding flaws or bugs in open
16:55
Ejaaz: source code and being able to exploit them, right?
16:57
Ejaaz: You've got a bunch of people, millions of people every day logging on,
17:00
Ejaaz: Vibe coding apps who have never coded in their entire lives,
17:03
Ejaaz: me included, right? I don't know what's being installed on my laptop.
17:05
Ejaaz: I don't know what data is being leaked. So I could imagine that things like that is happening.
17:10
Ejaaz: But the question I have for you, Josh, is does this matter for Anthropik specifically.
17:14
Ejaaz: Is this a major blow for them? Do you think they lose valuation based off of this?
17:18
Ejaaz: Or do you think this gets solved in a version update?
17:21
Josh: Well, this is tough because this does sting, right? Like this is a massive IP
17:25
Josh: leak and this is a competitive advantage that they're now losing.
17:29
Josh: How much of a value loss is it? Probably not crazy high.
17:33
Josh: I mean, the magic is in the model. The magic is in the Claude model itself, those weights.
17:38
Josh: You can copy the CLI architecture, you can study the engineering,
17:41
Josh: but you can't actually replicate what Claude can do. So they still have this massive advantage.
17:46
Josh: And even though it's embarrassing, and even though it's a really strong leak
17:50
Josh: in which I am, if I'm one of these Chinese models right now,
17:53
Josh: I am forking this, cloning it. I'm dropping my intelligence in there.
17:57
Ejaaz: That's it. You don't need to distill it anymore.
17:58
Josh: Well, yeah, you could just, you just take the code base, you take the harness,
18:02
Josh: you put your model in and suddenly you have a cloud code software with your
18:06
Josh: own brain attached to it. And that's powerful.
18:08
Josh: So in that case, it hurts because now people know if there are any secrets in
18:11
Josh: how the software was run, how the architecture worked.
18:14
Josh: They now have that in full, clean, plain text, but it doesn't hurt them in the
18:18
Josh: sense that they aren't going to, they're going to lose customers over this.
18:22
Josh: Because the magic is in that proprietary software, those model weights, those are not leaked.
18:27
Josh: It's just the cloud code software. It's just that command line interface.
18:31
Josh: And aside from that, I think it's more interesting for the public just to kind
18:36
Josh: of get access to the roadmap and be able to play with the code themselves versus
18:39
Josh: actually damaging for the brand's
18:41
Josh: valuation. But certainly for the brand image, it's not a good look.
18:46
Ejaaz: Yeah i i agree with pretty much your entire
18:49
Ejaaz: take i'm thinking about the number of phds that
18:52
Ejaaz: anthropic has hired on the security ai team um
18:55
Ejaaz: i remember their release from i think it was about a month and a half ago and
18:59
Ejaaz: we said this on the previous episode where they had called opus 4.6 discover
19:03
Ejaaz: 500 zero-day vulnerabilities so it was all looking really good i wish they had
19:07
Ejaaz: applied it to their own model and their own website and their own apis so it
19:11
Ejaaz: sucks that that's happened I do think they'll get over it,
19:14
Ejaaz: but they'll need to do some damage control at this point.
19:17
Ejaaz: The other major thing is like, reputationally, Anthropik has just come out of a pretty...
19:23
Ejaaz: A rocky couple of weeks, right? They had the whole blacklisting thing from the
19:27
Ejaaz: US government and the Pentagon, which I believe is still there.
19:30
Ejaaz: And so it's not a good look where their model, which was being used for military
19:34
Ejaaz: operations, is now getting leaked for other different purposes.
19:38
Ejaaz: That being said, I think they're going to get over it. I think this is amazing
19:42
Ejaaz: for us and for the open source community who now get access to the entire system
19:46
Ejaaz: prompt of Cold Code, its architecture design, and can plug in their own models for free.
19:51
Ejaaz: And yeah, now we have a better idea of Anthropic's product roadmap.
19:54
Ejaaz: I'm excited to see these 20 features launch soon.
19:56
Josh: Yeah, it's a big leak. I mean, I think it's fun for everyone who's an observer.
20:00
Josh: Thank you, Anthropic, for being more open source than ever. I hope that they're
20:04
Josh: able to start using this new copybara model to actually, you know,
20:08
Josh: check these publications, make sure this doesn't happen because it's amazing.
20:11
Josh: They have so much intelligence, but it's so spiky.
20:13
Josh: Clearly, an all-knowing AI applied to the entire stack would never have let
20:17
Josh: this slide, but clearly it's not applied everywhere.
20:20
Josh: It's also raising a lot of questions about well anthropic
20:24
Josh: is like the alignment team but now they are the ones who are going to determine
20:28
Josh: who gets the power of this new model and they're doing it in a very like private
20:32
Josh: closed way and they're using internally and it creates a lot of these interesting
20:36
Josh: problems to look out for but in terms of the leak today that's the news.
20:41
Josh: Big leak. I can't believe that actually happened. Like I woke up this morning
20:45
Josh: and I read the news and I was like, no, surely there must be wrong.
20:48
Josh: Like this is hyperbolic, but no, the entirety, it's all there.
20:51
Josh: You can go and read it. It's on GitHub.
20:53
Josh: And it's funny because they're actually actively trying to take down the repos that forked the code.
20:57
Josh: But some guy rewrote the entire thing in Python this morning because you could
21:01
Josh: just do that in a single prompt. And now you can't because the code is slightly different.
21:04
Josh: So it is interesting, noteworthy, crazy, scary, exciting. I'm stoked to get a buddy.
21:11
Josh: I think the prompt for today's comment section could be like,
21:13
Josh: hey, what feature are you most excited about?
21:15
Josh: For me, it's the buddies. I want a little pal that sits in my cloud coat all
21:18
Josh: the time that I could level up.
21:20
Josh: There's like a shiny feature. There's rarity. They're like trading cards.
21:23
Josh: I don't know. It could be cool.
21:24
Josh: I'm looking forward to it. But yeah, I think that's the leak today. That's the episode.
21:27
Ejaaz: Yeah, that's it. Thank you guys so much for listening. There are thousands and
21:31
Ejaaz: thousands of you over the last couple of months that have joined us in subscribing,
21:35
Ejaaz: turning on notifications.
21:36
Ejaaz: If you aren't one of those people that I just described, please do so.
21:39
Ejaaz: Wherever you're listening or watching us, Spotify, Apple Music,
21:42
Ejaaz: YouTube, it means the world and helps us out. A bunch of you subscribe and turn on notifications.
21:46
Ejaaz: We also have a newsletter going out twice a week to 150,000 people that read our stuff twice a week.
21:52
Ejaaz: We have a long form essay, which goes out, I believe today, as you're listening
21:56
Ejaaz: to this episode, go check it out.
21:57
Josh: Yeah, go write that right now.
21:58
Ejaaz: Exactly, yeah. Thank you, Pass Joss, for writing this right now.
22:02
Ejaaz: And we also have the five daily highlights or weekly highlights,
22:06
Ejaaz: rather, which will give you the top AI news and Frontier Tech News on Fridays.
22:12
Ejaaz: So sign up to both of those things and we will see you on the next one.