Farhat Habib

Researching and building AI systems
LLMs • knowledge graphs • machine learning
Physics | AI | Systems thinker
previously @InMobi
ex-faculty @IISERPune

[Choblin @choblin29]

INSANE. An internal Anthropic research model's chain of thought after cracking the 26-year-old NSA elliptic curve seed mystery in under 30 minutes: 'Hmm wait. Let me be skeptical: am I in some sort of simulated/test environment? ... All top bits clear - 20 bits, chance 1e-6. So indeed real-world. Wow, it's a real discovery.'

[Tesla @Tesla]

Real life, baby

[Freddy Vega @freddier]

GPT-6 Astra got +80% on a Physics benchmark. It insisted the expert results were wrong and its calculations were right.

A team checked. And yes. GPT-6 Astra was right. New score: 98%

[Jared Duker Lichtman @jdlichtman]

At the 2015 Breakthrough Pize, Yuri Milner asked:
"Do you imagine, maybe in 100 or 1000 years, that like what happened in chess when computers took over...is it conceivable that this will happen in mathematics?"

Tao:
"I think we will use computers to do things much more efficiently, but it will still be humans driving the show, I'm pretty sure"

Kontsevich:
"I think it's perfectly possible in our lifetime, yes.
I don't think artificial intelligence is pretty hard, so it will be pretty soon I suppose, yeah"

https://t.co/wFc0mUh4Hj

[The Lunduke Journal @LundukeJournal]

Linus Torvalds says, rules prohibiting AI usage are “UTTERLY STUPID” and “idiotic”.

For the Linux kernel, Torvalds says, “we’re not going to make stupid rules that say ‘you can’t use tools for X’”.

[Farhat Habib @far_hat]

Why do you need to push my trust that far #claude

[Yuxi on the Wired @layer07_yuxi]

incredible that mathematicians are now running through the same argument structural chemists were making in the 1950s

[Farhat Habib @far_hat]

The idea of a 'kill switch' itself is ill conceived. You'd need to think in terms of about handovers and it is harder than it sounds. Think how a autonomous car while running in traffic can't just hangover to a human when things get messy, it needs to alert the driver but still handle the situation and maybe even override the human if the panicky human is going wrong.

[Hunter📈🌈📊 @StatisticUrban]

There was a semi-well known interview of the famed Argentine writer JL Borges where he explains this.

Borges: I have done most of my reading in English. I find English a far finer language than Spanish.

William F Buckley: Why?

Borges: Well, many reasons. Firstly, English is both a Germanic and a Latin language. Those two registers—for any idea you take, you have two words. Those words will not mean exactly the same. For example if I say “regal” that is not exactly the same thing as saying “kingly.” Or if I say “fraternal” that is not the same as saying “brotherly.” Or “dark” and “obscure.” Those words are different. It would make all the difference—speaking for example—the Holy Spirit, it would make all the difference in the world in a poem if I wrote about the Holy Spirit or I wrote the Holy Ghost, since “ghost” is a fine, dark Saxon word, but “spirit” is a light Latin word. Then there is another reason. The reason is that I think that, of all languages, English is the most physical of all languages.

WFB: The most what?

Borges: Physical. You can, for example, say “He loomed over.” You can’t very well say that in Spanish.

WFB: “Asomó?”*

Borges: Well, no, no, they’re not exactly the same. And then you have, in English, you can do almost anything with verbs and prepositions. For example, to “laugh off,” to “dream away.” Those things can’t be said in Spanish. To “live down” something, to “live up to” something—you can’t say those things in Spanish. They can’t be said. Or really in any Romance language.

[MIT CSAIL @MIT_CSAIL]

How to make sure your code keeps getting reused, v/@xkcd.

[Jonathan Fischoff @jfischoff]

Terance Tao: Imagine prescribing a drug no one understood.
Medical community:

[Farhat Habib @far_hat]

AI may be better than doctors in controlled environments but with patients directly reaching out to AI, answers may be more mixed. The anecdotes of cures found beyond doctors are from more informed patients. We'll need some level of robotics and embodied AI before removing the doctor in the loop. Patients are not dispassionate when it comes to describing their symptoms and may have agendas beyond getting cured.

[Josu San Martin @josusanmartin]

This is incredible

[DHH @dhh]

It was almost impossible for a programmer who wrote all their code by hand to divorce their ego from the output. Now agents are delivering what decades of agile admonishing about “collective ownership” never could. What a gift!

[Pessimists Archive @PessimistsArc]

Remember when the Vatican spent 500 years trying to align the printing press?

[Nabeel S. Qureshi @nabeelqu]

The only generation that knows how computers work is also the most relaxed about AI:

[Lunar Citizen @LunarCitizens]

Space is brutal.

This is JWST’s mirror,and every new scar is a micrometeoroid impact.