chmod775

> It took TypeScript a year to go from announcing tsgo to releasing TypeScript 7.0.

These are not comparable at all. The bun "rewrite" really is more of a translation of a software that is mainly dogfooded, created with an at best loose regard for a wider ecosystem.

TypeScript 7.0. by comparison is not just a translation from one language to another. It's a true rewrite that at the same time has to consider a massive existing ecosystem (which needs to catch up first or you risk fragmenting it). Plus you might as well consider 6.x part of the effort, since that was the stepping-stone.

Also where bun really is just a runtime implemented with a bunch of glue code plus supporting tooling, TypeScript is much more complex. Change one part of TypeScript and you can get cascading issues somewhere else in the compiler easily. Change something in Bun and you'll likely just break some discrete part.

"Thought is language" is a folk intuition that has been proven wrong over and over. People with severe global or agrammatic aphasia still can score higher on intelligence tasks than you'd expect - even though this data generally comes from people who suffered a stroke, which generally isn't neatly contained to one's language capability. Also, more importantly, even in people with full brain function, the language network appears remarkably unresponsive under neuroimaging when performing arithmetic, logic puzzles, or theory-of-mind tasks. Also famously music tasks: Vissarion Shebalin lost most of his language capabilities to a stroke and still completed his fifth symphony.

There's correlation, and your language network likely augments your intelligence, but as proven by millions of animals, unfortunate humans and also some less-unfortunate human infants, you really don't need language for intelligence.

Where language helps most strongly is metacognition: evidence suggests that it is severely limited without language, and I suppose that is where the common belief that thought is language comes from: the moment you try to think about your thoughts, you use language!

My theory (and this with literally no evidence) is that we use the language network for metacognition precisely because it is not that involved in primary thought. Important to note that "we" here means humans: animals appear to demonstrate metacognition even without language.

Birth control is older than the Bible[1], it just wasn't as effective. Humanity as a whole was very aware how babies come to be.

Ways to have fun without coitus aren't a recent invention either.

Human reproduction and raising children is an unusually taxing, time-, and resource-intensive process, which is why we've applied our considerable intellect to plan for that since forever.

[1] https://en.wikipedia.org/wiki/History_of_birth_control

One reason our ancestors had plenty of children is that children were a large net-benefit to them.

Many modern societies are structured in a way that makes children an economic negative to their parents. From an individuals' POV, the "smart" play is not having children and hope others do.

Individuals believe that they do not need children because they have saved up for retirement, not accounting for the fact that if everyone did this and there was nobody left to do any work once they reach retirement age, that money will be worthless and they'll be unable to retire. The alternative is mass-scale immigration, and that's becoming unpopular. For an example of what happens if a society has neither enough children, nor immigration, see South Korea. Poverty among the elderly is rampant and they are increasingly being forced back into the labor market.

Many western societies needed to start applying a large multiplier to parental benefits decades ago.

Most omnivores and carnivores won't touch humans with a ten foot pole either.

But that's because they're either too smart to try or we have altered their genetic makeup by culling anyone who strayed too close.

It's language and speech patterns that seem designed to trick readers into believing that claims are correct, even when the claims aren't based on anything and are possibly wrong.

It was rewarded for this during training for some reason.

Alternative theory:

The LLMs only way to "think" about abstract concepts is through language, and this leaks into into conversation it has with humans.

But humans generally prefer to communicate on low levels of abstraction, through a back and forth, until the hard-to-express higher abstraction exists in the head of everyone involved - without ever being directly communicated. This is because we don't think using language. Language is merely a lossy translation of our thought into something expressible, happening after the fact or alongside it.

So when the LLM starts speaking to us using patterns and terms it created for itself during training to encode abstract thought in language, communicating with it becomes painful.

They don't exactly say. I was extrapolating from some sessions usage (my figure was per month, so weeks times ~4).

Many models are much cheaper through their subscriptions' included usage. That could be what's happening here.

Claude gives you something like $5000 of tokens on a $200 plan.

> Illustrative of course: the point is that "for every human" is not far out there given how crazy powerful hardware has become.

> There are not very many systems one could conceive of that could do useful work 1/s for ~every human on the planet.

Roughly a hundred servers on 100 GbE ports each could handle one small UDP packet per human every second, and would have a plenty of cycles and RAM per human left to actually do some work.

That fits in a shipping container.

Illustrative of course: the point is that "for every human" is not far out there given how crazy powerful hardware has become.

The more interesting finding is that it's still the second most expensive model (after Fable 5) by a long shot.

At least two models (GPT-5.6, Kimi K3) match its score (~1-2% diff) for half the cost.

> Codeberg is not for me, I could never get over them keeping the anubis anime girls on

To be fair, that seems like a bit of an odd detail to obsess over.

A service could be bright pink for all I care* as long as it just works without me having to worry about checking their blog for the latest decrees being handed down on a daily basis.

* Take that with a grain of salt because my - in some people's eyes - questionable stylistic choices are the direct cause of a couple hundred userstyles and scripts out there.

> Are you consciously choosing to completely ignore

Yes. Let me stop you right there: Yes, absolutely. Whatever they claim their reasons are, it's not my problem. I don't care.

The last Codeberg story that popped up here just this week was something about them banning Bitcoin. At that pace you should worry whether you're next.

Which is why my advice is: Avoid the headache. Go somewhere else.

Keeping up with this stuff is for people who define their identity through virtue signalling and whatever personality trait they have decided to make the problem of everyone around them.

Let them have their little corner of the internet and just stay away. There's plenty of alternatives run by people who don't feel the need to bring their personal baggage into a cooperative environment.

> i don't perceive any intelligence drop at 700-900K tokens

I have countless examples of this for Fable 5 and Opus 4.8 at just half that many tokens. Sometimes they can't even remember the correct workflow even if they're merely an orchestrator and have been going through the same 5 step loop multiple times for pretty much their entire context window.

Wall of shame with an Opus and a Fable oopsie from just yesterday: https://i.imgur.com/znSJ3OE.png and https://i.imgur.com/DLOHQ1B.png

One of them is particularly egregious because it uses a CLI tool to step through the stages that literally tells it what to do next. How do you fuck this up? Needless to say that tool was revised to force the agents to acknowledge everything happened, but I'm sure it will find a way to eventually mess up anyhow.

And orchestration should be the easy part. It's pretty much just writing and adjusting workflow files. Don't get me started on code quality - but I realize that's subjective and my bar is unusually high. There's a reason I have them run adversarial verifiers along multiple axis before their output is even worth my attention.

The judge explicitly ruled that training on the books was fine.

They just should have bought them, rather than pirating them.

Also LLM output is not IP (in itself) in the first place, nor would Anthropic want to claim it is and that they have rights to it - that would drive paying customers away.

The issue comes down to at most ToS violations.

Issue being that if you took everything that's "probably harmless" at once, cramming your body full of supplements, you might just end up harming yourself with substances that are toxic above certain amounts (acute iron toxicity etc.). Also the more random stuff you take, the more likely it is that something you took actually does turn out to have been harmful later.

So no, you should not take something merely because it's "probably harmless".

> there's an uncensored model that you can run locally with llama.cpp

Correction: There's tens of thousands of them. They're easy to create, which is why everyone publishes their own.

Just put "uncensored", "abliterated", or "heretic" into search on huggingface/ollama/etc and pick any them. Fair warning: most aren't very good, essentially lobotomized, and totally broken if you enable thinking.

Also ignores the common vector of the maintainer getting compromised and their credentials immediately being used to push malware.

Most maintainer are going to notice that themselves.

Bonus points for also covering misspellings like "pregbant".

If it ever had the slightest chance of being useful on Earth, it would not be called a "galactic algorithm". In this case that algorithm might be more than galactic: there isn't enough matter in the observable universe to build a conventional computer able to execute it.

In fact a computer executing it would be so large that speed of light would be the main constraint limiting execution speed, not theoretical algorithmic complexity that ignores data locality.

Windows automatically tries to download and install drivers for some hardware you plug in, including monitors. That's what is happening here.

Here's an article explaining why one should care: https://privacypros.com.au/privacy-hub/articles/dv-safe-phon...

The tl;dr is that you can either share this data by accident through some sort of "locate my family" app, or because your abuser gets access to your Google/Apple account (for instance because you're signed in on another device they have access to).

The threat model here can be: domestic abuse victim flees a situation at home in a hurry, stays signed in on a computer. Abuser uses the sign-in on that computer to track their phone, figures out they're staying at their aunt's place.

Yes, you can avoid this on a regular Google phone as well, but that requires correctly configuring it (and a lot, such as location and search history, can be re-enabled remotely!). If you're running Graphene you are protected by default, rather than compromised by default.

This is more of an ad, not a review, and reads like the author has hardly any experience with the things he's trying out. That Z Image Turbo diffusion model would've also run on many consumer GPUs and with way higher performance for a fraction of the price. Misleading.

Let's ask the thing whose mistakes I iron out all day to think for me.

> The standard answer is greed: rapacious ambulance operators, owned by villainous private equity firms, exploit patients at their most helpless. But I don’t think that’s actually what’s going on. Ambulance providers are chronically unprofitable businesses; margins are thin, crews are underpaid, and operators exit the industry every year.

The author does not understand how private equity extracts money. The high-liability and heavily scrutinized business is intentionally left with little profit: the actual profits are funneled up the supply chain. This is why private equity buys "nonprofit" hospitals - they can now control who that nonprofit buys services and equipment from.

> A team I talked to recently wired up an agent to do something simple: pull a metrics API every morning, reshape the JSON, and drop the result into a table. Clean idea. It worked on day one.

So a team of people with access to "intelligent" help did not immediately figure out that they could just have the model write a small script that does the job perfectly every time? That's "call a doctor" level of stupid that should be impossible in vivo. I'll believe a lot, but not this. The rest of the article is not much better - only an overexcitable LLM could believe anything there constitutes a deep insight worth writing down.

Do we need a new category for this sort of made-up, probably LLM-generated, garbage? Is this AI brainrot blogspam engineered to be upvoted by people who don't read the contents?

How do you not die of shame publishing something like this on your company blog?

The application is cool, but there is little novelty here. All of the employed techniques are well-established.

I suggest removing "novel" from the title unless you wish to seriously disappoint some people.

> I wonder how these will fare in Saxony. I presume it's an industry which will attract and depend on highly qualified foreign workers.

Saxony has had a decently-sized or large (by German standards) microelectronics sector for decades: https://en.wikipedia.org/wiki/Silicon_Saxony

As you'd probably guess, they're mostly situated around Leipzig and Dresden.