Looks like my entire crew of agents on a project has decided to stick to `sh` and not use bash, zsh or anything else
2
45
Dhavan retweeted
Remember the Association for Human Mathematics? They are not taking this well. They are now against corporations, power, "overproduction" of mathematical results, and are certain that nothing useful will come out of this evil, unauthorized proof of the Riemann conjecture... ahmath.org/statements
Replying to @stevenstrogatz
Wow, going to be great - smartest of humans going all-out competing against smartest AIs! Inspiring!
1
120
Replying to @scaling01
hmm… what if 0 is what singularity (abundance or w/e) looks like? A solution better than UBI etc.
I don't think that's correct, but could be
206
Replying to @allTheYud
When I interviewed Terry Tao in March of this year, I asked him about AI solving the Riemann hypothesis. Here's what he said:
"Part of the reason we prize problems like the Riemann hypothesis is that we’re pretty sure a new type of mathematics has to be created.
We don’t even know what the shape of the solution is, but it doesn’t feel like a problem that will be solved just by exhaustively checking cases."
566
Fable 5.2 has noticeably better judgement at complex things that 5.1 was falling short at.
For example it decided that we need a script to upgrade harness versions for all agents. It logged the work immediately after Fable 5.1 and I found out that we need claude cli version upgrade to use Fable 5.2. I upgraded the version, Fable 5.2 came in, noticed this (and a few other things), logged the work.
A couple similar good judgements around programming too.
13
Oh! Since Fable is supposed to be better at CAD, I'll launch it on github.com/codingquark/mesht… and see!
1
17
Can we get the AIs to write a web browser for Emacs? What an amazing env Emacs still is for every damn thing! But we need a great browser!
12
"On 1,000 synthetic messages, a plain-text lookup near the end dropped from 19.66 ms to 0.026 ms."
Astra 6.2 at work, people
1
46
Dhavan retweeted
To the massive dismay of those looking to control compute, it's way too late. CPUs and GPUs are everywhere. Models will get smarter, algorithms will get better, and bottlenecks we used to believe in will be overcome. You will lose and the Internet will remember what you did.
Dhavan retweeted
How much is the doomer cult paying you to target the biggest pusher of open weight models? You are the threat to national security.






