Pyro

✊🟥🌹 Transhumanist 👤🔁🤖
Choo Choo Choo Choo Choo Choo Choo

[Pyro @PyrosAI]

https://t.co/YIkjjSVa0y

[Pyro @PyrosAI]

@theo @edzitron @edels0n You mean like he tried to formalize how GPT-4o was the absolute limit of what LLMs can reach?

[Pyro @PyrosAI]

On a serious note, at the beginning of all this, I really envisioned something like an adapter marketplace.

Imagine being Terence Tao, working with ChatGPT for years and gradually putting so much of your own knowledge and data into it that you can press a button in the UI to build a "Tao LoRA" with incredible math expertise. Then other people could pay a few bucks to use that "Tao LoRA" through the marketplace.

It's basically shareable, semi-online fine-tuning with a business model attached, and I still think you could absolutely make that work.

But instead we got GPTs... lol.

[Pyro @PyrosAI]

@zeitonline Haha, ganz gewiss.

[Pyro @PyrosAI]

@thdxr a big chart with company logos obv

[Pyro @PyrosAI]

@0XKageX0 @JensenHuang I don't want to imagine Jensen gooning at anyone ;_;

[François Chollet @fchollet]

I don't read/write code these days, I only instruct an LRM. But it's not because I think LRM code quality is perfect, or even good. And it's not because I think my instructions are always being perfectly acted upon -- I know they aren't.

The thing is: LRMs don't just write your code, they unlock new ways to control what you're doing and maintain understanding of your system, which are about as effective as reading code. You can better test your codebase, you can audit specific components, you can spin up visualizations, you can ask the LRM to red team, etc. You can do all of these things at high speed -- you couldn't before. The benefits of reading/writing code can now be achieved in new ways.

All in all, the ROI of writing code by hand is no longer looking good, not because LRMs got perfect but because they're so fast at doing things that you can develop new workflows around them, workflows that end up being more productive than the old workflows.

[Pyro @PyrosAI]

@kimmonismus yes got the same

[Pyro @PyrosAI]

@perrymetzger Imagining the Big Yud in an orgy did more damage to my well-being than AI ever has or ever will.

[Pyro @PyrosAI]

@perrymetzger A proper banger

[Pyro @PyrosAI]

@cgarciae88 Imagine calling an anime waifu AI song "super human propaganda"

[Pyro @PyrosAI]

@EMostaque bro it's an anime waifu music track. pls stop your bs.

[Pyro @PyrosAI]

@AestherML @Seltaa_ I can't wait to discuss Gödel and Schopenhauer with my toaster

[Pyro @PyrosAI]

@perrymetzger @Noahpinion And that work includes fooling all the precursor tracking, monitoring systems, and guardrails that are already in place today, so terrorists can't just kidnap some bleeding-edge virologist and get their supervirus.

[Pyro @PyrosAI]

@edzitron Thanks, OpenAI, for burning through lots of cash so my workplace can pay a few hundred bucks a month to enable me to watch anime all day while pressing Enter in Codex does 90% of my job. And Ed will try to tell you this is something bad... somehow...

[Pyro @PyrosAI]

@mark_k @GoogleDeepMind The amount of google "leaks" i have seen the last 6month. If it is not released it does not exist.

[Pyro @PyrosAI]

@rfleury yeah won't even take "a couple of years"

[Pyro @PyrosAI]

@tszzl oh noe you forgot your /s and now 90% of twitter user's neocortex region of Tegmark IV will go broke

[Pyro @PyrosAI]

@augusteprompt @rileybrown someone got paid to write these. lol.

[Pyro @PyrosAI]

@scaling01 They also wrote some of the best AI papers, so no surprise