Funding to pay the core team (via revenue/grants/VC) requires a lot of leadership attention for any independent company that is developing an open-source project as its main activity. Yet more leadership attention goes into other administration (Taxes/hiring/legal/policies/etc.).

I don't have any direct context, though I have run an open-source business (Zulip) for the last decade wearing both the CEO and technical lead hats.

But my simulation is that the Bun leadership team might well be spending 2x as much of their time working on the technology than they reasonably could have as an independent venture-funded company, just because they don't have to do all that other stuff anymore. (There's of course probably a significant bias in that focus towards whatever Anthropic needs from Bun, only some of which other users may care about).

So I agree. Personally, I would not be concerned unless you see the tell-tale signs of the team being reassigned to other priorities at the buyer, which tends to be obvious, because, say, the GitHub project activity falls off a cliff.

It's interesting how quickly people buy the "abuse" line of thinking. We understood (and knew for a long time) that the large AI labs are not monetarily profiting from subscription users that make heavy use of their subscription. That is independent of which agent/harness is used. The fair/real price for profitable use is the pay per use token pricing.

These labs play the game of trying to kill competition in the harness game (because third party harnesses risk commoditizing the underlying LLMs once they are all good enough), while playing a game of chicken with each other how long they can burn money that way before they have to give up.

At some point they have to price their product fairly, and the only hope they have is to have killed all competition by then, which is of course a game that they seem to be loosing. Useful models are getting smaller and cheaper to run every year and it has hit a threshold at which we will see continued development of third party harnesses even without the userbase of subscription users.

Basically the prime bet that they made (that one needs extremely expensive hardware to have useful AI) has already failed. The secondary bet that they can lock users into their ecosystem (which requires them to subsidize their harness via unprofitable subscriptions burning their capital) and be able to monetize that later will also fail. They will have to compete on merit alone, and that is much less profitable.

> Before the acquisition, Bun had to figure out how to monetize at some point.

I think it is insane that people got into a situation where they had committed to a javascript runtime that had to "figure out how to monetize at some point". It is also bizarre that some people are still hopeful despite it being acquired by one of the most enormously unprofitable companies in the most enormously unprofitable sectors of our industry.

Node supports Querying sqlite with tagged template literals.

https://nodejs.org/api/sqlite.html#databasecreatetagstoremax...

Yes, I've tried to run it in production on a couple projects. I had to back out from bun to node on both. One, there were huge memory leaks like you mentioned. The other, there were API differences that threw errors, in TextDecoderStream and such. Decided I won't try again until bun v2.

There's a GitHub issue for the freeze thing. Their security scanner passes the full dep list as CLI arguments, large monorepo on Linux and you blow past ARG_MAX. Spawn silently hangs, no error, --ignore-scripts doesn't help because the scanner is separate from postinstall. Been broken since 1.3.5 at least.

Additionally, with Typescript compiler rewrite, it is even less relevant.

This is cool! But AFAIK bun promises to be a one-stop-shop for all your JS/TS dev needs, while Perry is "just" a compiler from Typescript to native executables.

> it's a stretch to assume this will also translate into making Bun worse

For me it's far from a stretch, in fact it matches closely a pattern that I've seen repeated many times over at this point.

...try `bun repl` in your terminal

otherwise, bun has a big "batteries included" thing going on.

For instance,

- Bun.$ to run shell commands

- an entire redis client at Bun.redis

There are dozens of other examples like this

For rapid prototyping, complex glue scripts, etc. it's an absolute joy to work with. There is often no reason to pull in any dependencies to accomplish what you want.

If I had to guess, it comes down to speed of iteration. Claude Code is built on JavaScript, so Bun aligns well with their current stack.

Switching to Go or Rust would only make sense if performance were the main priority, which doesn’t seem to be the case. Their current setup lets them ship quickly. A rewrite in Go would likely slow that down.

Codex moved to Rust, and you can see the trade-off. Performance improved, but release velocity dropped. They’re also still catching up to Claude Code, so they don’t face the same pressure to ship as fast.

My guess: JavaScript runs in the Browser as well as on the OS. That way you can train a model to be able to interact with both fairly simple. You can also see that their harness, claude-code is also written in js. So I guess they are quite invested in that language anyway.

Claude is still better at writing JS than it is Go.

Yeah, it's the same pattern you saw in the early react days where open source devs would try to "woo" the react core team into getting recognition to sell consulting services or courses.

The bun people likely have some fucked up incestial business relationship with some >dev manager at Anthropic and the same pattern is repeating. Only this cycle it's going straight to acquisitions, which honestly seems like a worse strategy and Anthropic will def can the bun engineers in less than <3 years or whenever they face an actual budget crunch that they can't stave off with more gulf money.

I’m wondering why Anthropic, who has “the most powerful, hold me bro, AI in the world” just didn’t vibe code their own, better, version of bun? haven’t Dario said that coding is cooked in 6 month, like 12 months ago?

Acquisitions in this industry tend to lead to a certain inevitable conclusion. The software that has been acquired gets worse as the original team members cash out and their culture is replaced with the culture of the new owner.

Perhaps Bun will be the exception, but you can't say that the concern is unfounded.

The CEO of Anthropic has a habit of making outlandish predictions about how AI is so very close to replacing human programmers. Anthropic has been applying this belief to Claude Code and it has become a giant heap of unmaintainable spaghetti.

I think velocity is a real risk to stability, dogfooding or not. That's what made me swear off the python transformers library. It's doubtful that LLMs will change that calculus for the better.

Hasn't your team shrunk a lot? Word on the street is that many of Bun's employees left or let go in the time leading up to the acquisition. How many people are left working on Bun?

Has development velocity increased because you are merging large quantities of unreviewed LLM generated code? If so, I would be very worried about future stability if I used Bun.

The best feature Bun delivered recently is portable binary. That portability is a huge deal to me as my users are often on ancient Linux distros. Thank you. Both node and deno require recent Linux, more exactly, recent glibc.

Hey Jarred, first of all, thanks. I've been doing backend JS since the first release of node and bun is genuinely the first really big improvement in terms of DX. It's an absolute delight to build glue and scripts with... Bun.* just seems to have everything I need. Bun.$ is revolutionary. etc. etc. I'm hoping to run a collection of backend services on it in the near future but it seems the general consensus is that there are still some gremlins holding it back (memory leaks, etc.)

Can you shed a little light on the recent giant rust based commits though? Are you guys moving away from zig? These kind of big curious movements and the spectre of giant LLM-based commits are not exactly confidence inspiring.

Why not just write a small helper library to add back the features you miss? Node includes SQLite and Argon2 at least, if the issue is the interface then that is easily fixed.

Node supports auto loading .env and also supports sqlite

When did Node add native TypeScript? Can you run "node main.ts" directly without any dependencies?

To be fair, Node didn't have any of these things until Deno & Bun challenged it. Deno didn't seem to move the needle by itself very much for whatever reason, but Bun's existence has had a tangible effect on the Node Technical Steering Committee. I would even argue that much of the current impetus has been driven by Jarred Sumner's savvy social media marketing. It got people talking, and Node is better because of it.

Additionally, Bun's push for covering as much of the Node API as possible has pushed Deno towards the same level of compatibility, and now most code is basically runtime agnostic. I'm not sure if I'll ever actually use Bun in production, but I'm glad it exists because the JavaScript ecosystem has been much improved simply due to its existence.

I would mention deno as the main competitor

Looks like AI slop

https://github.com/PerryTS/perry/issues/139

I disagree with the overall premise: Before the acquisition, GitHub had to figure out how to monetize at some point. Now, even though their parent company does some shitty practices with their other software (Embrace, Extend, Extinguish, MS Windows), it's a stretch to assume this will also translate into making GitHub worse: Being worried makes sense but I remain optimistic about GitHub.

> Now, even though their parent company does some shitty practices with their other software (claude code), it's a stretch to assume this will also translate into making Bun worse: Being worried makes sense but I remain optimistic about Bun.

Anthropic acquired Bun for their own benefit, to protect and grow their investment in Claude Code. Not for the benefit the JavaScript community at large. Sounds obvious but I guess that has to be pointed out. Outcomes will follow incentives in the long run.

You might be underestimating the effect that corporate policies and culture have on the product.

Some teams have a push now to go all in on AI; don't even look at the code. I've seen this in action and the results are probably what you'd expect. Works great at some level, but as complexity accumulates (especially across a team with different "technical vocabularies"), the end result is compounding complexity and mistakes and no person or team knows how the software actually works.

No human testing of software or QA; unit + integration + give AI control over the browser/tool. Yes, this how some teams are moving forward now. So some of this may be that Anthropic's culture will end up causing shifts in how the Bun team operates and thinks.

If this type of culture and mindset becomes the norm, I think either the models have to get a lot better or the software quality is going to decline.

Matt Pocock has a great talk here: https://youtu.be/v4F1gFy-hqg

    "Code is not cheap. Bad code is the most expensive it's ever been. Because if you have a codebase that's hard to change, you're not able to take advantage of all of the bounty that AI can offer.  Because AI in a good codebase actually does really, really well."

Once bad code starts to compound on itself, it's going to be really hard to break out of it.

Just 3 days later from the blog post, a branch with a potential vibe rewrite from zig to rust surfaced in the bun repo

> Now, even though their parent company does some shitty practices with their other software (claude code), it's a stretch to assume this will also translate into making Bun worse: Being worried makes sense but I remain optimistic about Bun.

Can you point to any examples of a company with shitty practices buying one without shitty practices that didn't end up with the shitty practices diffusing through the newly-acquired company within a couple of years?

This is a good take, and I hope you're right.

One favorable way to phrase it for Anthropic is they acquired Bun because CC and other internal tooling depended on it so heavily and they questioned it's future as purely OSS.

It remains to be seen how things will actually unfold.

There's a VC behind that too.

Is Claude better with Javascript than it is with Go code? Seems like it could be true.

I doubt those SWEs could have used anything other than JS.

One of them is a much more efficient but obscure programming language from a competitor, the other is what the web is built on.

Saying that you “work on Bun” is such a radical understatement. I have my reservations about Anthropic, but I don’t see how Bun could go wrong with you at the helm. And I’m sure that you are putting the stability and funding of a larger organization to good use :)

I’ve been a Bun maximalist since the beginning. Thank you Jarred!!!

"I have a vague concern, so I'm now using a shittier toolchain. You shouldn't do it though." is a weird post format.

I dont think so, but recent release includes a terminal markdown renderer built-in which means, even if handy, most of the focus is to make Claud Code great. I am not worried though, at least no yet.

These are just my opinions man :)

I like Deno because there is no "install" step for users, you just run it.

Started using Bun last year for some quick tests and it ended up fully replacing Node for any new projects after using it for over a decade.

I've reduced my dependencies 5-10x. Got full TS and JSX/TSX support with zero setup. Watch mode is instant. You can deploy a single binary.

I kept waiting for all the breaking issues people complain online but my experience has been nothing but positive.

> Both claim better performance, but that only seems true in cherrypicked benchmarks, and in my tests (granted about a year ago at this point) both alternatives under-performed Node in my workloads.

1) You need to retest again, mainly because Bun's own native tools should be faster than Node's.

2) My experience is the opposite: For the niche uses I'm on, the rendering process is done 2-3x faster with only a few changes to use Bun's tools.

> Deno is just Node with a restrictive, annoying permission system

I find Deno's permission system amazing! (although I didn't stick with it until v2)

Everything is closed by default but you're able to write code like normal.

Whenever it needs a permission the code pauses (like `debugger;`) and the terminal asks you "hey, should this script have access to this file/folder"?

- You say yes and the code continues (no need for exceptions).

- You say no and the code stops.

Then after your program has run, you put only the answers you said yes to in a deno.json file and it never has to ask again.

---------------------------------------

I'm currently working on a project that takes in heap of files from one one set of devs, processes them with a heap of files from another set of devs, then compiles and outputs the final product.

The file structure goes like this:

1. Group one devs

2. Group two devs

3. Build output

4. Compiler

So group one only works in their folder, and group two only works in their folder, but needs to see group one's folder.

With Deno it's stupidly easy to do stuff like:

- Scripts in group one only have file read access to group one.

- Scripts in group two only have file read access to group one and two.

- Scripts in the compiler only have file read access to group one and two's folders, only have file write access to build-output folder, and can read the env file in the project's root directory.

- One specific file is only allowed to access a specific URL and port

- Another specific file is only allowed to use the FFI to access a specific shared object.

I don't need to worry about a dev's script accidentally using the wrong file because they messed up the path.

I don't need to worry about a dev accidentally overwriting a file and losing data.

I don't need to worry about a dev blindly going down the wrong road because an LLM convinced them to.

I don't need to worry about a dev using LLMs agents that are trying to make the project do something it's not supposed to do.

I don't need to worry about a dev including a dependency that's doing what it shouldn't be doing.

I don't need to worry about the equivalent of `rm -rf ./$BUILD-OUTPUT` but the env file wasn't set up correctly and $BUILD-OUTPUT is empty/undefined evaluating to `rm -rf ./` and nuking the project's root.

I don't need to worry about supply-chain attacks.

I don't need to worry about namesquatting attacks.

There's so many things I don't need to worry about.

It's such a breath of fresh air.

It's just: you guys read from here, other guys read from here, the compiler writes to here.

Whenever something doesn't fit, the program stops and tells you what file is trying to access what permission.

---------------------------------------

aside: Node added a permission system but it's completely broken by design. Everything's open and you have to manually close each permission yourself. Oh, you don't want this project to have file write permissions? Lets just turn off the file write permissions (and forget to also turn off the subprocess permissions to spawn a shell which rm -rf's the wrong folder).

Agreed- node/npm are exceptionally well run and designed. Personally I also prefer the non-TypeScriptiness.

Given the complaints about Firefox and Safari not adopting Chrome OS Platform APIs, and shipping Chrome all over the place, I am not sure about people standing on the ground and ethical considerations.

> where people are more conscious of ethical concerns and willing to stand on their ground to avoid repeating past mistakes

Would be interested to hear what makes you say that. I don't see anyone being conscious of ethical concerns more than they were before. I can see slightly more BDS people, for example, but outside of that not much.

Claude did write me a simple wrapper so I can keep using tagged template literals in the same way with Node.

Reminds me of the back and forth competition between Node.js and io.js that we had to endure back in the day. Worked out for the best in the end.

v22.18 promoted type stripping from experimental

Seriously. This baffling omission undermines the credibility of the whole thing.

"I want a serious Node.js alternative."

So you ignore the one developed by the same guy?

I personally switched back to vscode as I started using Claude and Opencode more for the AI flow, and I didn't see much added value any longer. Also, I was incredibly frustrated that they decided to hide the close button and finally, there were weird issues with editor groups spawning at unwanted times. They might be able to fix it, but I felt that they were starting to reach the limits of what you can do with a "live fork".

The main complaint about Cursor I see online is that it's expensive.

Otherwise if you are looking for and IDE first approach with great AI integration it's the best product out there. I prefer it over CC/Codex.

I'm no staff+, but the tooling (package manager and TypeScript runner) is very fast.

bun run is <1s for my projects, while watching for file changes. So the iteration speed is quite pleasant.

Google seems profoundly uninterested in the agentic coding world though. gemini-cli is underwhelming, Antigravity not super compelling, and the Gemini model itself absolutely terrible and non-competitive in basic tool use necessary for coding, even inside their own harnesses.

It's fine for other purposes though. Which are arguably a much larger and lucrative market.

That's assuming Google can outpace Nvidia, which may or never will. Nvidia is not just going to sleep on it.

Doesn't seem that bad if you're convinced they're the only viable market dominator.

Their third part harness move seems like more than incompetence.

Yeah, I'm none too happy with anthropic right now, but what's happening to Claude code is just your typical garden variety mismanagement of a project that grew way too fast for its owners to reasonably handle.

If you were right Anthropic's ARR would be going down but it's not. They just surpassed $30B up from $14B two months ago.

> We understood (and knew for a long time) that the large AI labs are not monetarily profiting from subscription users that make heavy use of their subscription.

"profit" is a weird concept in the software business. it might be true that there is an opportunity cost to these users, either because they displace other potential users by using up capacity, or because they would be willing to pay more if forced. but I don't believe that anyone is losing money on inference costs on any of their plans.

> At some point they have to price their product fairly

they are competing in a market. if most of their costs were inference then this would be a good thing, because everyone would have roughly the same prices, so as long as they had the best model they would win. in fact model development costs eclipse the cost of inference, and is something that non frontier labs get for much cheaper by distilling from the frontier companies.

> They will have to compete on merit alone, and that is much less profitable.

that's not really true. google won search on merit alone, and were massively successful as a result. the trick is that everyone from the poorest shmuck to the richest businessman uses google, so they win through scale. in ai, google and openai are making a bet that they can do the same thing. there's only really room for one winner at this game, even two is stretching it, so anthropic has to win by being the smartest model that only high end businesses use. that's a very risky bet.

>Basically the prime bet that they made (that one needs extremely expensive hardware to have useful AI) has already failed.

Honestly, I don't think it's that cut and dry. Their bet is that the marginal utility of having a smarter model more than makes up for the cost of the additional high-end hardware.

And honestly, if you look at their frankly insane revenue growth since Opus 4.5 released, they were right.

>The secondary bet that they can lock users into their ecosystem (which requires them to subsidize their harness via unprofitable subscriptions burning their capital) and be able to monetize that later will also fail.

I think we're already past this point, honestly. They lowered usage limits, blocked OpenClaw then tried to remove Claude Code from the $20/mo plan. They have always had low market share for the consumer chatbot market and don't seem to care about catching up to OpenAI there.

> Useful models are getting smaller and cheaper to run every year and it has hit a threshold at which we will see continued development of third party harnesses even without the userbase of subscription users.

As of May 2026, how much money do I need to spend to buy hardware to have a local model that is 80% as good as SOTA services for assisting me in writing code?

As for that 80%, how many minutes per LOC will I be waiting, and how many attempts per query will I be wasting while I wait for it to come up with something sensible?

What about the data they are accumulating, for non-training purposes? That data isn't of negligible value; the "subscription cost" is really a "harvesting data" opportunity. Don't be naive to that our data is not incredibly valuable.

The thing is, the harness _is_ the model at the end of the day:

https://en.wikipedia.org/wiki/Turtles_all_the_way_down

The problem is this.

If not VC funding, then what? Volunteer work? So other people can make money off it?

Our industry has no answer how to fund infrastructure.

You've got FAANG companies using open source projects built by volunteers and doing meagre grants every once in a while, not nearly enough to pay a SWE salary. A smattering of hard to get grants from NLnet, etc. And then places like Anthropic or Grok or OpenAI "buying" open source teams to pull them inside, which inevitably leads to drama.

I don't know what the answer is, but there's a serious issue here. Similar situations in the 80s were why the FSF was founded and the GPL established. (Not to fund, but to protect the rights of authors and users)

I switched to Deno because it is the only option out of the 3 that allow monorepo workflow without building .d.ts files. Bun and Node both do type stripping or compiling of TS, but it only works for the entry package of the running script, not any of the linked dependencies from the same repo.

There are still things I dislike about Deno, but it really does make package development a lot simpler. JSR is a great upgrade from NPM, and Deno makes it so simple to publish to both NPM and JSR. Strict IO permission system and WebGPU support are also nice to have.

> wrap a project into an `*.exe`

Deno makes this simple too. Though that's where it's bundling features stop. Honestly I am okay with that, I'd rather use Rolldown or Vite for web or library bundling.

The built-in sqlite and testing functionality is the reason I started using it over pnpm/Node.

> Bun has no REPL

Bun has a really nice REPL, can recommend https://bun.com/docs/runtime/repl

Deno has sqlite: https://docs.deno.com/examples/sqlite/

The Bun DX is infinitely better than Node's, especially for Typescript projects

Sure, but at least on Node 22 I think I have to pass `--env-file=.env` option to make it pick up .env.

No disagreement, but this article was posted 2 days ago, the argument isn't relevant right now.

January of last year. Yes.

https://nodejs.org/en/blog/release/v23.6.0

Personally I much prefer Deno as it's also doing a lot more work to unify the backend and frontend JS APIs.

I would too ... but not as the winning competitor.

For their first year two of existence, bun tried to do npm, but better. For the first year or two of their existence, Deno tried to reinvent npm.

The key result is that after that first year or two Deno had to walk back their decisions, to create a Node-ecosystem-compatible tool .. and as a result, they're now significantly behind bun (at least by all metrics I've seen).

> Good question, and you're basically right — let me show the smoking gun.

:vomit:

I would mention Node as the main competitor. It isn't moving as vast as the VC-backed ecosystems are but its future is a lot more assured.

the AI replies itt are cringe

Wild example as it has been corrupted by VC money as well. I wouldn’t touch deno either.

I envy your experience. Its driving me crazy on a near daily basis now.

Using that logic, Node.js is just a wrapper over V8.

There is the SpaceX acquisition rumor, but that's not why.

I only use Cursor through the CLI, and while the UX of the CLI is pretty bad, I've found their harness (the prompts they use and orchestration of LLMs) to be nothing short of incredible. I can't comment on their agent development environment given I haven't spent a lot of time with it.

The reason I'm moving away from Cursor is cost. Unfortunately, if you want to use the SOTA models from both OpenAI and Anthropic you basically have to go direct through their subsidized plans.

just conjecture but possibly because of the rumored acquisition plans from SpaceX (that's why i stopped using it)

Disagree, you definitely don’t want to be looking back saying “hm I knew it, I saw the signs, should have trusted myself”

Plus it’s not a huge lift right now

If Disney acquired your favorite IP, wouldn't you get worried?

Simplicity.

bun file.ts

And it’s been this way for years.

Don’t care about what’s in a package.json file or if there is one. Can do this without tsconfig file as well.

Do you use Gentoo as OS?

> Basically the prime bet that they made (that one needs extremely expensive hardware to have useful AI) has already failed.

I thought the prime bet was that the winning lab who reaches takeoff through recursive self improvement will make a galactic superintelligence. Not saying I believe this but the people running the labs do. Under this scenario if you are a few months behind at the pivotal time you might as well not exist at all.

It's a big leap to go from "some users may be using large quantities of tokens" to "the labs are burning money on subs in an attempt to kill the competition."

Lots of businesses have subscription programs in which a small number of users are money losers, but which in aggregate make money.

It's not even obvious that the labs are losing a lot of money on even a minority of users; the rate use caps are fairly aggressive for Anthropic, and a cursory analysis of likely actual cost of serving tokens shows they are high margin products at the API level and unlikely to be unprofitable within the usage constraints provided to subscribers.

I do think subscription models make commercial sense because users want predictable costs, and it's a club good in which marginal token cost is zero which helps consolidate their customers' purchasing volume to one provider. But that's a different claim than them serving it unprofitably to kill competition.

Also, they (Anthropic) are transitioning many of their enterprise customers to API consumption billing anyway.

> We understood (and knew for a long time) that the large AI labs are not monetarily profiting from subscription users that make heavy use of their subscription.

I dont think this is "understood" or "known" to anyone except Ed Zitron. Subscription plans like Claude Code also have rolling usage limits, it could be profitable. Inference is very cheap and unless you're using OpenClaw no one is actually maxing out the usage window at all times. I'm sure in aggregate the subs are not money furnaces.

> These labs play the game of trying to kill competition in the harness game

Anthropic and Google are arguably playing that game. OpenAI's Codex CLI is open source and entirely optional for use of the GPT Codex models.

I partially agree with you, but I also think that it's good that people can make something they want, that seems to have no monetization path, and have some hope of being bailed out.

It's not great that the search for profit will usually corrupt projects, but the other most common option is that the projects don't exist at all. It's very rare (or it used to be before this year) that someone can do something like this on their own with no compensation. So now at least Bun exists.

I know people say it is unprofitable but I wonder if there is a way to verify it is truly is. I will not say any details but I worked for a giant company which was barely making money YoY but somehow the bonuses for heads were bigger and bigger given a proxy metric related to profit.

There are way too many ways companies arrange to pay themselves and never be profitable to avoid taxes.

Perhaps runtimes used to make companies profit shouldn't be free if we're not funding those services in a way we do other public services (taxes)?

I never thought I would come around to the Java way of thinking on this, but companies are abusing the public good.

Deno has been great for wrapping the dozens of REST API's I need to use in the world in MCP. The no compilation thing means that I can push and it's literally deployed in seconds. I run several dozen of the little servers for various use cases, it's a very cheap way to build an automatable life

I am gonna check my sources better next time lmao, sorry!

So does node, since v22: https://nodejs.org/api/sqlite.html

They even added sql template string queries like recent popular libraries in v24.

I just built a project using it.

Ah my mistake, this wasn't the case last I used it, thanks for pointing this out, I checked briefly and referenced stale data.

I don’t believe so, Go has simple rules, snd in my experience Claude is excellent at writing all the boilerplate needed

Problem with Go is the type system is rudimentary, so you can't "restrict" AIs as well as you could in Typescript.

I honestly did not know Node had all of these things now, but that is great to hear. Clearly I haven't been keeping up well enough.

Node didn't have all of these features when I initially went down the path of choosing Bun, so I have a number of existing projects that have Bun baked into them.

Yes, same for me. I was skeptical at first but things have really improved over the past 2 years

There is also 1 line to load .env in node without passing an argument

I know, early deno was rough and frustrating. But it is now _the_ main competitor to Bun. What makes you say it is behind? Are you talking about features or usage?

The source code of Claude Code and Gemini CLI contradict that.

Node has both built-in sqlite and testing functionality. Lots of reasons to like bun! But these two are interesting ones...

In what particular way? I've been using Typescript a lot more recently (unfortunately XD) and I've found the native experience in Node to be totally fine.

Ironic that this comment is in thread advocating for usage of Go:

"The key point here is our programmers are Googlers, they’re not researchers. They’re typically, fairly young, fresh out of school, probably learned Java, maybe learned C or C++, probably learned Python. They’re not capable of understanding a brilliant language but we want to use them to build good software. So, the language that we give them has to be easy for them to understand and easy to adopt. – Rob Pike"

In what world is Go an obscure programming language??

Perhaps it could go wrong because he uses AI robots to generate responses to issues on claude code that are also generated by AI robots? Just bots talking to each other like moltbook. It shows a level of AI maximalism that is absurd, concerning, and funny. But probably par for the course for someone working at Anthropic. I can imagine being surrounded by people doing similarly foolish things only encourages the foolery.

Certainly true but there’s a line between “better DX” and… well, kinda laziness

That is type stripping and is incompatible with syntax that requires transpilation, so it is not native TypeScript support.

Freedom from the NPM mess was why I started my project from the ground up in Deno in the first place.

Is there any Typescript that is compiled wrongly? If so, go ahead and show me.

Ok, but Node is the status quo. As replacement to the node runtime bun and deno are the two contenders at the moment

Are they wrong though? AI or not, compiling code is math - not philosophy. So what's wrong here?

Corrupted how?

ah ok, yeah that would give me pause as well.

I agree with your assessment that the harness is incredible and so I get a ton of mileage out of Auto + Composer 2. This is my workhorse.

Admittedly, with Opus 4.6+, GPT 5.5 I just haven't used them much and as I gain more experience I can see what the hype is all about. But to me, the answer isn't $200 max plan, it's bifurcating the work. Call me a spendthrift!

Yes, but I wouldn't stop watching the originals. And I wouldn't stop watching anything until I saw what came out.

How is that better than Deno?

Also this works on node too now, don't sleep on node improvements

Really? This was like a intern level response lol

I don't think this race to superintelligence idea should be taken too seriously. It is great for headlines and get peoples imaginations up. It is mostly a marketing gag.

I look at superintelligence this way: software engineering used to be considered amoung the most mentally demanding jobs one can have. And in this field more and more people give up large parts of their job and become approximately product managers to let the machine do the engineering part. So we are about there. Who cares that there are some puzzles in some "synthetic" benchmark in which humans outsmart AIs?

That's just what they told the gullible investors to get money.

OpenAI just has more runway and has convinced its investors that it is as much about hardware (stargate) as it is about anything else. So they think they can/have to afford keeping the software side more open to not make themselves look stupid. Google is more of a down to earth company with other business to lose and isn't bought into it as much.

It's a bit insane, but the cost of switching to regular NodeJS is low (for all but most bun-specific projects).

All valid points though, I'm pessimistic about Anthropic still actively diverting resources to these side quests when tough times hit (which might be in a week for all we know).

I think you have some nostalgia about Github's stability before the acquisition.

This looks like a vanity project: the value gained switching from zig to Rust is likely to be negative at best, without even the usual caveat devs use of "learned a new skill".

Anthropic bought actual engineers to undo the slop their vibe-coders produce with reckless abandon: https://x.com/jarredsumner/status/2026497606575398987

However, these engineers, too, now start to vibe-code with reckless abandon https://x.com/jarredsumner/status/2048434628248359284 and https://x.com/jarredsumner/status/2049780223311548729

This is a joke, right?

1 9 of uptime later

I am asuming you are European as I (I live in the Netherlands) have not seen any outage of GitHub as well...

you can own your upstream supply chain while simultaneously being less responsive to user pain points

Own your supply chain. Reduces risk.

> Anthropic acquired Bun for their own benefit, to protect and grow their investment in Claude Code.

I’m unclear about this. What’s the business case? I use Gemini CLI a lot, which runs on Node, and I can’t see anything that would be improved by using a different JS runtime. It’s not something you notice as a user. Node is mature, stable, and perfectly fit for the purpose.

If Anthropic were public and if these decisions were comprehensible to the average investor, an acquisition like this ought to cause the stock to plummet. Luckily for the people involved, there are no constraints like that in the current market.

You dropped this: </sarcasm>

I'm with you... I think it's helped Node.js a lot to have Bun and Deno implementing new features that help push node forward. I think it's been a bit of a miss not integrating npm into node along the way... Mostly in that npm is a separate org from node, which is its' own issue... I kind of like JSR a lot myself, so hope it continues to pick up some traction.

It has eject functionality and all the subcomponents are independent open source projects.

Seriously. It's one thing to use AI to write code but spamming machine generated garbage when talking to another person is just rude.

What do you mean by corrupted?

Been using Deno happily for many years now.

Did I miss anything?

But are you saying the harness is driving you insane? Or the model? Because Bun is,only the harness, and that part has been improving over time if you stay on the stable channel

I'm still getting pretty good code out of it, but I only use it on side projects. Is the issue with their odd limit system?

Indeed it is

Genuine question: why not just wait?

If Bun stays great, you saved yourself some time for switching, and got to keep using Bun.

If Bun worsens, you spend the same time for switching, just moved a bit later, and got to use Bun for a little longer.

I did for over a decade, but it does not go far enough with supply chain security.

I bootstrapped a new generation of Linux distribution from 180 bytes of human readable x86 machine code all the way up.

https://stagex.tools

only if said galactic superintelligence takes immediate steps to kill all its potential competitors, or hoover up all the world's resources, or some other aggressively zero sum thing. otherwise I don't see what difference it makes down the line of you have the second superintelligence rather than the first.

and that's under the assumption that you can create a superintelligence that will continue to slavishly serve your agenda rather than establishing and following its own goals.

One thing I don’t understand about this viewpoint (which I understand isn’t your own): why does one benefit so tremendously from getting there a month before competitors? I’m sure having a month of superintelligence with no competition would be lucrative, but do they think achieving superintelligence first will impede competitors from also achieving it a month later?

I work in the video AI world.

We gave up on subscriptions long ago. They're rinky dink and get you a paltry amount of utilization before they run out.

The per day per seat costs can exceed $1000. This is already normal for studios, and it's already producing positive ROI.

There's simply no way to price video any other way than by usage. I suspect the same will come for everything.

Then explain why they started banning all third party harnesses, including those that work through Claude Code, if it still makes them money. They are cutting off profit for no good reason?

I think there were reasons to doubt that heavy subscription users are unprofitable before they did that. OpenClaw was just the tip of the iceberg.

Why don't they make token pricing dynamic if that was the case? It should then allow heavy user to get even more for their money than with the current subscription model where they can't adjust to current infra availability.

It may be that "in aggregate" sub users are (not yet) a loosing business. But in all fairness, the more useful AI gets, the more it will be used. And the more it will be used, the harder it will be to make subs cheaper than token pricing. The only counter-weight are new light users, but those will also become heavy users over time, the more useful it will be for them. And at some point it will be hard to onboard light users in the first place, because the laggards will require even more intelligence and value to get them over.

Are there any situations you would compare this to historically?

To me, the obvious comparison seems to be Docker. Their tooling revolutionized software development and made cgroups and containerization accessible to the masses. Yet they generally seem to have failed to extract payment from users, even with managed service opportunities.

It seems to me that there are substantial obstacles to monetizing a project licensed with even a weaker OSS license like MIT. I think this is especially true for projects that don’t have managed service / “open core” potential.

Any gratis project you rely on runs the risk that it will no longer be provided gratis. That alone is not a strong basis for making decisions.

I'm not the parent poster which is why I still stick to looking at the people...

If you start seeing the people that created bun leaving Anthropic, then I'd probably start to worry. And I haven't seen any sign of that yet.

> I think it is insane that people got into a situation where they had committed to a javascript runtime that had to "figure out how to monetize at some point".

Why? What's the risk? It's open source. Also, speaking of open source, we are happy to commit to open source projects that have no monetization, nor any plans to ever monetize.

Bun is not a "product" at Anthropic though, it's a tool for its developers to build products. IMO as long as it remains that way, the incentives for its developers will remain fairly aligned with the incentives of people who use it outside the company.

A good example is React. Facebook's interest is that React be performant (website performance is correlated with time spent on said website), reliable (also correlated to time spent), quick to build on (features ship faster) and popular (helps new recruits hit the ground running). That's fairly well aligned with what developers outside of Facebook want too.

Sure, since Facebook's server is written in Hack it means we'll never get a truly full-stack React, and instead we'll need third parties for the back-end (Next.js, Tanstack Start, etc). But Facebook building react also means it will always be someone's job to make sure this Framework works well in codebases with millions of modules.

This is all independent of any shitty practices with their other software. And this has been for decades at this point.

I don't disagree with the notion, but what is up with the dev community championing influencers that work no real jobs and just sell courses where they reread the docs to you at $500 a pop (this gent, $1k a pop)?

"Profitable" is the wrong metric, really, it's whether it is sustainable - can development continue indefinitely given the current financial situation?

I hope Node eventually gets a WebSocket server like Bun has.

Well sure, I can find you hundreds of dead ends in mutations, but when you start parsing through what a harness does, eventually it'll just be another more deterministic and constrained model to do less creative things.

Yeah they should have used Rust ;)

The world of many, many companies.

It's a little heartbreaking. The DX of Bun is legitimately amazing and perhaps even revolutionary (I say this as a long-time javascript backender). I'm all for LLM-based development velocity enhancement, but it does really feel like they are taking it too far and moving too fast.

I don't really see the issue here. They are language models after-all and they work by talking. Whether it's one model talking to itself (i.e thinking/reasoning), or one model talking to another it amounts to the same thing.

Yeah I don't want it to silently read magic config files

TypeScript is a wide umbrella. For instance, Experimental Decorators are shunned by many (including me), but they are still used by millions. If I don't use any syntax that requires transpilation, am I not still using TypeScript?

Now that we have `satisfies` and `as const`, there's really no reason to ever use an enum. In my opinion, TypeScript is best when it is simply used as Language Server, and it should never have had runtime implications in the first place.

Node's built-in profiler doesn't work with Typescript, which is one part of Node not natively supporting TS. Idk how it is in Bun, cause that made me abandon TS rather than abandoning Node.

In 2026 world, whilst you're advocating for 1980.

Node v22.7.0 added support for TypeScript syntax that requires transformation:

`node --experimental-transform-types example.ts`

As for whether this matches your definition of "native support" or not...

Source: https://nodejs.org/en/blog/release/v22.7.0

Isn't that mostly just enums?

Is there anything else that doesn't run as valid JS if you strip the types (and maybe some other extra keywords)out?

Genuine question, in my head there's not much, but TS has a few weird corners I maybe haven't used

Modern Typescript does not need runtime features.

Its the product overall, and its impossible to say where the issues are but I tend to think not the model since the changes seem to be able to occur overnight. So likely a combo of harness and service-layer.

Nothing, Deno raised capital back in 22

https://deno.com/blog/series-a

You should probably caveat any post you make about security concerns with that, so people can more easily judge whether your concerns line up with their threat model.

What happens to every new project you start on bun and have to switch?

Great!

Anthropic/OpenAI aren't planning to have their superintelligence take over the world, but they're still afraid that someone else will do it.

Well no because no one is going to be coming in to work building the next AI model after the Singularity.

We’ll all be bblbrvkxn46?/4!gfbxf’mgv5fhxtgcsgjcucz to buvtcibycuvinovrYdyvuctYcrzuvhxh gcuch7…:!

It's a tenet of the eschatology from the singularity ideology that was developed on online forums over the last few decades.

The viewpoint is baked into those assumptions and boils down to the power of exponentials and poor application of game theory.

Assuming it can't super hack all computer systems and cripple competing SI incubation to at least increase its lead time indefinetly.

The assumption would be that in the lead time it has the super intelligence at least takes a small lead and undermines any paths a later arriving super intelligence could take to interfere with it's goals, which naturally includes stopping competing SIs from becoming more powerful in a way that could undermine it.

So assuming the super intelligence has goals and work towards them it will be initially trying to solidify its own power, iterating on that small lead, assuming it's the smartest super intelligence[1], should be enough to win. The scary part is that assuming no guardrails [2] it's going to be as ruthless as possible in achieving those goals. That does not necessarily mean it will appear ruthless in achieving those goals, just as ruthless as it judges optimal.

1. Which being so smart one of it's chores would have been reinvestment in making itself smarter than competition and being smarter than its makers has a good chance of actuating those self-improvements.

2. In the internal balancing of goals sense not the don't feed the mogwai after midnight sense.

> There's simply no way to price video any other way than by usage. I suspect the same will come for everything.

I don't think there's any way for all of the current AI models to work except as a usage model. The question is whether or not people are willing to pay for it that way in the long-term.

It sounds like it is producing positive ROI for your side, but I’m curious what the bean counters at the studios think of the bill when the budgets tighten.

If each additional user is a net benefit for them, but they're still struggling to find enough capacity, it would make sense to cut down usage from existing users so they can onboard new ones.

They're trying to capture the market! Can't do that if you have to stop onboarding users because NVIDIA are struggling to manufacture enough GPUs for you.

I feel like you're overestimating how much power the people who sell their company have to prevent anything that gets pushed from the top

I think parent commenter meant that what's insane is that js runtime is not treated as an utility which should never be monetized. It's as if GCC developers haven't figured out how to monetize, but they are willing to at some point.

6 month earnouts..... wait until july

I have followed a simple rule in my career, if you offer training/courses I don’t listen to anything you say.

I consider this a hard rule, like ad-blocking (this is exactly that, blocking ads as each talk is an ad (or ad in disguise).

I see the thread and the patience of the other guy to continue talking to the AI, it's impressive

Im on pay-per-use plans so its not the limits thats the issue directly, although the product development process could lead to them trying to fix limit issue and breaking the product as a whole.

The main issue is side effects of effort/thinking it seems. It hallucinates at a much higher rate and skips research in a ton of edge cases even with effort of MAX and disabling adaptive thinking, even on 4.6. Ive said before, but using opus today feels like using sonnet from ~October timeframe. Its not anywhere near what opus 4.5 in January felt like, or even opus 4.6 on release (notably 4.6 on release _really_ over-researched even simple tasks and that behavior is almost entirely gone now even with max effort so they are definitely re-tuning these things on the fly and degrading the experience as a result).

EDIT: I also have a very high suspicion that the way they hydrate thinking is buggy and/or lossy (or maybe unintentionally lossy which leads to bugs). So many behaviors just make no sense at the level I have my setup tuned (I have everything set to "just charge me the most money to hopefully get the best results") and the fact that I havent changed anything while using it daily for months and months on end, but have been getting worse and worse results.

Which of the two is responsible for it ignoring being in Plan mode and trying to implement shit instead of just writing up a plan?

This is also assuming that AGI is even possible. So far there is no evidence that this is actually doable over anything but billions of years (and even then we have no idea how nature really managed it).

Edit: Meant to say AGI (superintelligence didn't make sense). Superintelligence is undefinable at the moment so even considering if it's possible or not is more of a philosophical thing/si-fi thought experiment than anything else.

One could argue that AI has already started to hoover up all the world’s resources. AI buildout as a percent of GDP is already high and still rising.

If OpenAI has the second superintelligence they have to merge with the first and cooperate. It's a provision in their charter.

A week of superintelligence should be enough to take over the world, or at least sabotage your competitors. And even if someone else gets there a week later, they'll be permanently one week behind the curve (until the AI hits some physical limit, I suppose).

But that's all just sci-fi worldbuilding.

A month with a superintelligence at your hands could be quite impactful, especially if you're willing to break the law / normal operating decorum in the pursuit of protecting what you have. A superintelligence, if wielded so, could destroy your competitors in a great many ways, including the relatively-benign solution of outcompeting them, to exploiting them and tearing them apart from the inside.

A genuine superintelligence is a very, very scary thing to have under the control of one person or organisation.

positive ROI for customers?

It's a shame that VCs have corrupted a $200MM/year business into the perception as a failure. Who cares if the VCs didn't get a large return, or if the outsized impact of the software didn't quite fully capture the value created. $200MM/yr without aggressive R&D or operational costs could be an incredibly healthy business.

Maybe we should stop trying to build so many billion dollar/year businesses and work on more sustainable models.

The audio and 3D card pioneers in the PC world.

The ones that were first to market went all bankrupt, or were acquired by others that came later into the scene.

I'm not the biggest fan of the influencer community, but I think that it mostly boils down to many learners preferring video content over written material. I've gotten used to reading documentation now, but I remember it being extremely intimidating when I was first learning. It was nice to have someone break stuff down into simple terms for me.

To be fair to Matt Pocock, I know he worked for Vercel and Stately for a while before doing content full time. I can't say anything about his AI content, but I did some of his free lessons when I was learning TypeScript. They included interactive editor lessons and such, so it wasn't just empty videos and fluff like some of the influencers.

> Bun is not a "product" at Anthropic though, it's a tool for its developers to build products.

Doesn't that just make it even worse? If Anthropic can't even afford to spend the engineering effort on making sure their core product functions properly, why should we assume that they'll be investing serious resource into what is essentially some upper manager's loss-leader pet project?

If Anthropic is financially hurting, why shouldn't they put Bun on the bare minimum of life support?

I'm thinking about your comment... It put many wheels to spin...

Tldr; I think the don't care about what will happen to the company in medium or long term.

---

Are any of those companies looking for stability or sustainability?

I have the impression they are completely aware of the diminished return effects and they will explore the moment to the fullest of their capabilities promising even more absurd things when the results are even smaller.

I do agree there is a considerable improvement comparing from a year ago but definitely not ground shaking as it was from the year before to the last.

Many of the promises turns out to be empty or at least having huge number of asterisks to it.

I think there are flags everywhere. From minor things such as everyone using different benchmarks or plotting performance differences on weird choices os axis and ordering.

Other mild things such as promoting the "system" created a compiler from scratch when such compiler does not even do a hello world and runs and gave output binaries running 300x then the counterparts.

(I am aware there was a misusage of the agentic benchmark to build a compiler but there was an active choice on how to tell the story. Given other movements I am not quite sure if I believe it was an accident)

There are other red flags such as people rolling back to previous versions of models because they can't get the new one to work properly.

Other situations such as the affirmations that they have such "dangerous" model that apparently seems to be more of a benchmark trick than real results with <100B models being able to replicate the benchmark results only by changing the methodology.

I don't think we are yet in the turning point where everything will collapse but my feeling is that we are going in that direction unless something that makes these models much more intelligent AND efficient.

It makes sense to not hire a person when you can have a machine for the same job for the same price. But AI prices are getting higher than the returns do the margins for it to be a sensible choice are getting smaller.

That all said, I say again that I think that they are completely aware of this effect. Not because they understand the technology but because this happens more frequently than not. Because of this, I don't think they care to be sustainable. All of them,smell that they will take the money and leave the ship to sink.

Node supports ts natively now though.

In Node 26 it will be removed

https://github.com/nodejs/node/pull/61803

I'm not advocating for anything in this thread, and definitely not for Go's outdated language design, although Typescript is so unsound it isn't better in my book. But obscure is not synonym for regressive.

https://www.typescriptlang.org/tsconfig/#erasableSyntaxOnly covers them all, I strongly recommend running with that option enabled to be future-proof.

enums and decorators mainly. There are also subtleties such as having the ts file extension in imports. Also imports aren't transpiled in cjs so you need to need es modules.

I'm using it in my projects with no issues.

Your comment might lack explanation, but indeed the TS team has mentioned multiple times that they don't want to add any more features that require transpilation (as opposed to "dumb" type stripping and being a strict superset of JS).

IIRC they "almost" recommend against using them (the last part, I haven't researched again now).

But the usage of many features has reached a sort of point of no return, so I hope Node will go the route of making the experimental transpilation the default for TS files at some point.

Goes to show how strong the appeal of syntax is, especially enums.

To people coming from languages with enum support, it just looks so much more organized to use them, compared to union types, despite all of the (many) drawbacks.

Claude has definitely gotten stupider (even on the latest Opus).

I used to be able to give it certain commands, and reliably count on it to do the right thing. Lately I give it identical commands and it just starts doing something idiotic, instead of the correct thing (that it did 50 times prior).

To an earlier poster's point, it's probably the model, not the harness, and I understand Anthropic has to make money someday (and they're not now) ... but I'd rather see a visible doubling of price than a secret halving of the capabilities (which seems to be their current plan).

That approach is enshitification.

Yeah I have found worse results if I don't leave it on the highest setting. I have gotten by with Pro and a little overage buffer so far. I have found it working pretty well for what I'm using it for but I have really only been using it a couple months now.

The model I believe. Also a pet peeve of mine…

oh absolutely, no argument there, the case for AGI is pretty weak. I was just saying that I am even more sceptical that any of this is a "first or nothing" scenario - that is one of my biggest pet peeves about the entire tech sector.

I'm not sure anyone thinks their charter carries much weight at this point.

I haven’t followed Docker’s case in particular, but how much investment was required to get it to that point? If it’s a case of “How do you become a millionaire? Start as a billionaire and invest in Docker”, then the perception may have some basis.

>they'll be permanently one week behind the curve

What if the competitor's architecture is able to produce tokens twice as fast. What if the competitor secures a 1 month exclusivity deal on Nvidia's next generation?

I'm the "AI cringe guy" if you want to call it that. Yet I am still waiting for someone to produce typescript that compiles wrongly.

I have limited time, and the little feedback that guy provided turned out to be perfectly well answered by AI. So sorry, but either you actually criticize something actionable to just shut up, but I don't have the time to debate this if the simple few lines don't get answered.

> So far there is no evidence that this is actually doable over anything but billions of years (and even then we have no idea how nature really managed it).

"The brain is so mysterious and unique, that we should abandon all attempts to even try to apply results like the general approximation theorem to it and discard all signs that some approximation is happening."

Why we don't see signs of intelligence in the universe? The simplest self-replicator requires accidental synthesis of the sequence of 200 (or so) RNA nucleobases.

BTW, your argument could have been applied word-for-word to powered flight in 1899. In short, argumentum ad ignorantiam.

ASI is the acronym you’re looking for. It stands for Artificial Superintelligence.

Arguably it’s already here. ChatGPT knows more than any human who has ever lived. It can carry out millions of conversations at once. And it has better working memory (“context”) than humans. And it can speak and write code much faster than humans.

Humans still have some advantages: Specialists are smarter than chatgpt in most domains. We’re better at using imagination. We understand the physical world better. But it seems like we’re watching the gap close in real time. A few years ago chatgpt could barely program. Now you can give it complex prompts and it can write large, complex programs which mostly work. If you extrapolate forward, is there any good reason to think humans will retain a lead?

Don't blame machines for our folly. This is just standard bubble behavior.

If I interpret "a machine superintelligence" as "a classroom of 300IQ humans," I'm not really sure how this is true? You still have material and energy constraints, you can't think your way out of those.

No, look into his actual work history (sorry being a paid marketer isn't working as a dev). Was only a dev consultant for like two years before pivoting into full time influencer. Trust me, I know more about these types than any normal human should.

Because they need it to work, so that everything built on it works too.

Building developers sell you the apartment, not the elevator room, the electrical room, mechanical room, etc. They will make all sorts of controversial decisions with the apartments; odd layouts, ugly flooring, weird pricing, tacky finishes, etc. The "core product" is the money-maker, that's where the egos clash, priorities change, and where they try to charge as much as possible while they cut costs as much they can.

No one is buying the electrical room though. It just has to work. Yes, you'll make it as cheaply as possible; no flooring, no paint on the walls, no interior designer meetings to argue what's the right tone beige for the walls. But it'll do what it needs to do. It'll keep the lights on. Otherwise you can't sell any of the apartments.

Same thing with Facebook; there's active incentive to introduce all sorts of dark patterns over their app, to ignore certain bugs, to unnecessarily change things, etc. But none of those incentives are present with React. The incentive is to keep React reliable and performant, and to keep the team lean. I'm sure it's similar with Bun in Anthropic.

And to be clear, Anthropic definitely spends most of it's engineering effort making sure their core product "functions properly". This "functions properly" is just different for us as clients vs them as a corporation. There is high overlap, since they need to keep us clients happy. But a well-functioning product at a company is one that leads to money. I'm sure very capable engineers pushing the okrs they care about.

I think they're doing too much vibe coding and not enough QC... I don't think it's a matter of not having the resources so much as running while juggling multiple sets of scissors..

I think there's also an idea that sustainability requires constant maintenance and improvement; but TeX is sustainable and it's approaching a limit function - it (plain TeX) gets no new features, but has very few remaining bugs.

AI is already in commercials, TV, and movies. Companies for the most part just don't tell you because the reaction of the general public is "eww, AI".

It's already here in a big way. You just won't be told about it until the public lightens up on the "AI hate".

> but I think that it mostly boils down to many learners preferring video content over written material

99% of the times that's not learning, but productivity porn.

1. At least 99% of all species that ever lived on Earth are now extinct. I.e., that's life.

2. "But for a beautiful moment in time we created a lot of value for shareholders."

> It can carry out millions of conversations at once.

You're anthropomorphizing it, this isn't what it's doing. It's being fed a series of text and predicting what comes next the box has no context about the other "conversations" it's having and doesn't remember them.

We have more like Artificial Superstupidity.

Ultimately our current model is extremely unlikely to perform better than the sum of current human knowledge. Godlike super-intelligence is a pipe dream with the current LLM based approaches.

No. To realize the possibility of powered flight one only needs to look at birds. AGI, on the other hand, is another word for God.

Absolutely correct. It's not just smart to use AI in this way, it's efficient. And here's the thing most people don't get, you are saving time.

If you would like more insight, just say the word.

ChatGPT can only respond to a prompt, and in the context of that prompt. It has no continuous awareness of anything. That isn't superintelligence. We are easily fooled because we have stupid monkey brains.

What if that's just the mechanism the machines take over the world?

Natural selection doesn't care why something replicated a lot.

This. You can't learn by viewing, you learn by doing.

That bill is gonna come due at some point for "developers" leaning heavily on agents.

Failure for those species though.

For the concrete problem we're discussing, you can hack your competitors out of existence, replace all of your knowledge workers to shed costs, hyperoptimise your logistics, etc. It's not just intelligence, it's speed and scale.

Bostrom's Superintelligence (2014) is a bit of a dreary read, and I didn't finish it, but it pulls no punches about the leverage that a superintelligence might have in our highly-connected world.

I think the vagueness of statements like this is why a lot of people (myself included) are just so very skeptical. Surely some company wants to brag about their use. I don’t doubt it’s found its way into certain spaces, but by and large a lot of the “big” claims have been demonstrated to be borderline fraudulent. That Brad Pitt/Tom Cruise AI fight is fake. It is misleading. Taking existing green screen choreography and using AI to impose Brad Pitt and Tom Cruise’s faces is not what it is being sold as. Darren Aronofsky’s AI works are not good either. They can’t seem to hold a shot for more than a few seconds, why is that?

If the argument is that AI is being used in the background or for some VFX, sure, I’ll buy that. It’s just another tool, then. If it is being used to generate entire scenes, there’s no evidence of this, unless something like that atrocious holiday Coca-Cola commercial is a herald of our future.

As written, your claim is just handwavy. I get you might not be able to cite anything concrete due to NDAs or whatnot but, you also have to understand why a lot of people find this kinda unpersuasive.

Just define "general" as "as general as allowed by math, physics, and practical limitations." Or use a conventional reading of AGI as a human-level intelligence (which we, naturally, have a working example of).

I can respond directly to this, I’m a former VFX industry person and still fairly well connected.

The the former you suggested. Background plates and the like. The lack of actual creative direction tools, trite visual style, lack of consistency/repeatability and complete inability to be edited or adjusted easily make it a non-starter for most tasks. Compositors are fast, LLMs are slow at that scale. There are tools like ComfyUI that sit in the “we’re running experiments/useful sometimes” category.

Loads of ML tools are in use and incredibly handy, but fit into that tool category, but actual wholesale video/image generation is not that prevalent, no.

> For the concrete problem we're discussing, you can hack your competitors out of existence, replace all of your knowledge workers to shed costs, hyperoptimise your logistics, etc. It's not just intelligence, it's speed and scale.

For the concrete problem we're discussing, that hypothetical belongs in a Marvel movie, not reality. In the real world, you can't 'hack your competitors out of existence', and you'll be going to prison very quickly for trying this sort of thing.

I did say

> especially if you're willing to break the law / normal operating decorum

in my original post. If you have a superintelligence, you have something that can find and take advantage of every exploitation vector in parallel - technical, social, bureaucratic - and use that to destroy a company from the inside. A superintelligence that is subservient to its operator is an informational superweapon.

I agree that this sounds fanciful, but you can see what existing cyberattacks can do to organisations; it does not take that much imagination to gauge how much worse it could be when the process can be automated and scaled.

> A superintelligence that is subservient to its operator is an informational superweapon.

The five dollar wrench attack will put an end to that operator's use of an informational superweapon.

> I agree that this sounds fanciful, but you can see what existing cyberattacks can do to organisations

What can it do? Generally, a minor disruption to operations.

It consistently does a lot less than what law enforcement can do to you if you start messing with other rich peoples' money, while having enough of a presence to own a super-intelligence and a trillion-dollar data center.