dinobones

There are a lot of people working in software that have limited knowledge of information theory or machine learning.

Unfortunately LLMs are magic to them.

When you lack a first principles understanding of how something really works, you will bolt on hacks and duct tape on the wrong approach.

Not relevant to the article.

Also, that study is extremely dubious/barring on fan-fiction.

Homer (if he was even a person) recited oral epic poetry (that we've lost most of) and we've written down an extremely lossy version of it.

Why should we take any of the mythical character's interactions with mythical gods as any insight for the brains of real humans???

It'd be like year 4000 historians citing World of Warcraft claiming the Lich King is a very obvious example that humans lack an internal monologue because he was not able to distinguish his own motives/actions between Ner'zhul/Arthas (both spiritual entities inhabiting the "Lich King" body).

> Agents can oneshot games that are actually fun.

Stopped reading right there.

Is this really that big of a flaw in this implementation? I don't think it's worth the additional complexity to address it. (Encoding depth information in some way, trying to detect "flat" surfaces, whatever).

Discerning a camera taken image of an image is typically very very easy. The collors/exposure/etc will all be obviously wrong in ways to a human, even without doing any analysis.

This is a good product but the naming/branding is pretty unfortunate.

Typesafe.AI sounds like some typescript/structured output type of tool…

What even is “system one” ?

IMO the product/tech is really there, just needs better communication.

As someone with keratoconus, I hardly noticed! This is what all text looks like for me.

AI researchers: Yeah our AI models are a bioweapon threat and a nation state level cybersecurity threat and also a genius math researcher and very dAnGeRoUs!

The world: Cool, can you rewrite this email with a professional tone.

Why are you assuming there’s some taxpayer that will pay for my transportation?

I fully intend to pay my transportation costs myself.

It’s not about getting a free ride, it’s about being able to get to work while sleeping, saving money on housing, and pocketing the difference.

Here's another idea:

The places where people want to live have no land or archaic zoning laws.

I wouldn't want to live two hours away, but if I could sleep/work during my commute, and pay $400/month for it, I might do it. My housing costs would be like less than a third of living in the metro where I work.

Supercommuting becomes slightly more palatable.

This is like the “why not use FTP yourself” response to the YC Dropbox page.

Yes, but good, standardized developer tooling is valuable.

Postman could’ve also been a “just use CURL and dump to JSON” for example.

I actually am bullish on this. Tmux and ssh are somewhat clunky and probably under utilized by most people. I struggle to find commands, navigate sessions, see history, and so on.

Also in an “agent” based world where you might ask an agent to execute things for you, how do you manage that context well beyond just giving it its own Tmux session to drive?

It’s a good idea, and I’m surprised no one’s built it sooner.

Just start pricing in bytes input/output. This whole "token" and "tokenizer" thing is an implementation detail that shouldn't even be leaking out into the API.

Providers change tokenizers all the time with model updates, and it's often not even possible to query/figure out how text is tokenized without actually just sending the LLM a request.

Just switch to charging for bytes of intelligence. Please. Claude Shannon figured this out decades ago.

Don't underestimate corporate waste. If it's not someone's job to care for something, they really won't.

Even before this AI wave, it was common for me to see spinning dev environments for like $3k/month that hadn't been used in months on AWS.

Yeah, this implementation and their behavior these past few weeks is especially laughable when you consider that they consider themselves “philosopher programmers” or whatever.

You would think they’d be more reflective and introspective about these brash moral decisions. Their product quality is akin to my CS capstone lab group.

Those walkways are only in between each set of gates, there aren’t any actually near gates or anywhere near seating. Where did you sleep lol?