Maybe they should implement a graph based trust system:

You need your favourite academic gatekeeper (= thesis advisor) to vouch for you in order to be allowed to upload.

Then AI slop gets flagged and the shame spreads through the graph. And flaggings need to have evidence attached that can again be flagged.

The endorsement system already works along that line: https://info.arxiv.org/help/endorsement.html

It's probably not perfect but in practice, it seems to have been enough to get rid of the worst crackpotty spam.

They already had a basic form of this for a while [1]

> arXiv requires that users be endorsed before submitting their first paper to arXiv or a new category.

[1] https://info.arxiv.org/help/endorsement.html

I've often thought that similar trust systems would work well in social media, web search, etc., but I've never seen it implemented in a meaningful way. I wonder what I'm missing.

You mean like endorsement? https://info.arxiv.org/help/endorsement.html

Science reduced to people with a phd?

not a bad first order filter.

can you think of a better one?

The whole point of the scientific method was that we could ignore the source of the information, and were instead expected to focus on the value of the information based on supporting evidence (data).

If we go back to "Only people that have been inducted into the community can publish science" we're effectively saying that only the high priests can accrue knowledge.

I say this knowing full well that we have a massive problem in science on sorting the wheat from the chaff, have had so for a VERY long time, and AI is flooding the zone (thank you political commentator I despise) with absolute dross.

> anthropomorphism

I think it's a topic worthy of discussion. But I would propably not leave it to Searle...

serious question:

> no change in survival rates

> less series A

would this not imply that companies got more efficient at using their seed funding?

(But then again: The real dip in series A funding starts in 2018; so we might still see a dip in 10y survivability starting 2028)

> LLMs + harnesses can actually learn.

No. No, they don't

Fullstack lean when?

I like both worlds: Tinkering and vibe coding.

My shift in perspective is really: Not all code deserves to be hand-crafted. Some stuff can be wonky as long as it does it's job.

(And I think the wonkyness will reduce in vibe-coding as harnesses improve)

I think this is 100% the right direction:

Instead of imperatively letting the agents hammer your codebase into shape through a series of prompts, you declare your intent, observe the outcome and refine the spec.

The agents then serve as a control plane, carrying out the intent.

Very much agree. I like the imperative vs declarative angle you take here. Thank you!

Awesome!

A few questions:

- Is there a list of host languages?

- Can it live in the browser? (= is JS one of the host languages?)

The host is written in Rust, with `extern "C"`, which makes it able to be loaded as a C library by programs written in other languages. Most languages have support for this.

It's also designed to be run in an event loop. I've tested this with Bun's event loop that runs TypeScript. I haven't tried it with other async runtimes, but it should be doable.

As for the browser, I haven't tried it, but you might be able to compile it to WASM -- the async stuff would be the hardest part of that, I suspect. Could be cool!

A good rule of thumb:

- Don't even let dev machines access the infra directly (unless you're super early in a greenfield project): No local deploys, no SSH. Everything should go through either the pipeline or tools.

Why?

- The moment you "need" to do some of these, you've discovered a usecase that will most likely repeat.

- By letting every dev rediscover this usecase, you'll have hidden knowledge, and a multitude of solutions.

In conversation fragments:

- "... let me just quickly check if there's still enough disk space on the instance"

- "Hey Kat, could you get me the numbers again? I need them for a report." "sure, I'll run my script and send them to you in slack" "ah.. Could you also get them for last quarter? They're not in slack anymore"

> How many people can do that?

The answer is simple: By definition only about 100-300 people.

There's only 100 of the "worlds biggest companies" (assuming this refers to the top 100). And companies are usually started by 1-3 people.

Similarly: There's usually only 4 participants in the top 4 of a tournament bracket.

(The question is a bit: what does "can" even mean in this context and the answer im hinting at here: It's not individual skill that creates companies ex-nihilo. It's our economic system that produces companies.)