tlrobinson

I always assumed LLMs would be one component of “AGI”, but there would be “coprocessors” like logic engines or general purpose code interpreters that would be driven by code or data produced by LLMs just in time.

100%. I know I’m in the “get off my lawn” phase of my career when I see things like MCP and LangChain, but know I would have been excited about them earlier in my career.

Tailwind’s abuse of CSS classes as a DSL has always felt like a hack to me.

You shouldn’t need a special editor extension to get highlighting/autocomplete/etc, just use TypeScript and a CSS-in-TS solution.

Or not even imagery collection, but testing to see what our response is like. If we're scrambling and unable to explain/contain it, that's useful for an adversary to know a little about our current defensive capabilities.

This is all inevitable. At worst it's pulling the issues forward by a few months or years, and I don't think anyone will meaningfully address the problem until it's staring us in the face.

I believe the internet needs a distributed trust and reputation layer. I haven't fully thought through all the details, but:

- Some way to subscribe to fact checking providers of your choice.

- Some way to tie individuals' reputation to the things they post.

- Overlay those trust and reputation layers.

I want to see a score for every webpage, and be able to drill into what factored into that score, and any additional context people have provided (e.x. Community Notes).

There's a huge bootstrapping and incentive problem though. I think all the big players would need to work together to build this. Social media, legacy media companies, browsers, etc.

This also presupposes people actually care about the truth, which unfortunately doesn't always seem like the case.

> not responsible for the technical successes

He's not doing the nitty gritty engineering day-to-day, but he understands enough to ask the right questions, give his teams permission to try ideas that seem crazy at first, and sometimes come up with those ideas himself (e.x. supposedly catching Starship with "chopsticks" was his idea).

$50,000 income, 4 year product lifespan?

Obviously doesn’t apply to all engineers.

Who do you think writes those CAD and design tools that help “actual engineers” solve the same problems over and over?

I don’t think it’s at all unreasonable for an engineer using a device for 8+ hours every day to pay an additional, say, 0.5% of their income (assuming very conservatively $100,000 income after tax, $1,000 extra for a MacBook, 2 year product lifespan) for the best built laptop, best screen, and best OS.

LLMs are also often used in the search component of RAG, by generating embeddings that are then indexed and searched.

(I don’t know if that’s how Bing AI works)

I think in both cases some amount of interpretation was required:

- for Perl, OCR to a character set

- for Piet, manual “convert[ion] into a clean image file using close colours from the Piet palette”

(parent commenter is founder/CEO of Stability AI, Emad Mostaque, I assume)