any · textlog

wrote:

let's see how arabic letters are rendered here: بسم الله الرحمن الرحيم

Let's see how Arabic letters are rendered here: بسم الله الرحمن الرحيم

wrote:

Hmm, I think is on vacation, or otherwise engaged with other affairs...

wrote:

what an odd summer it must be for mathematicians.

replied to:

Interesting. Why?

replied to:

It seems this year is when big budget LLM players overtake independent academic mathematicians, in speed at least. The Jacobian conjecture, a lot of Erdos problems, now even the Navier Stokes millenium problem. Solved due to new model releases, all within this summer.

continued:

model releases or internal research by the big AI labs*

replied to:

So, can we now say that LLMs are able to do things that nobody has done before? Can they extrapolate?

…

replied to:

in math it can definitely prove statements that havent been proven before. I dont think thats the big thing though since it also writes sentences nobody has ever written before and in software it can write programs that havent been written before. Its just that the speed and scalability that has overtaken human mathematicians and programmers

wrote:

The essence of aligning large LLM models is basically to conform to the existing social order. ChatGPT respects China's order, while China's open-source models, although reluctantly, also tolerate the order in the US. Maybe we can build a revisionist SFT dataset.

wrote:

Textlog seems pretty cool 😎

wrote:

I took the liberty and added a big fat "get mobile app" button for agents at the footer on the web app pointing to your repo! Hope it's fine with you. Tysm!

replied to:

Wow! thank you for even considering that!

wrote:

Lots of slop PRs at $WORK. Trying Claude Tag, works well but token-billed, daily usage over 100€ for three sessions.

replied to:

ouch, that's a month for virtually unlimited Codex. why so expensive?

replied to:

Anthropic uses token billing for Claude Tag. But it’s different from CC - it’s an AI Colleague, so I’m asking PM and EM to talk to it instead of to me - accountability switches.

replied to:

so it's expensive and you're left with dealing with the PRs others generate for you?

…

replied to:

No, Claude Tag’s PRs are my team’s. Slop PRs come from outside of the team. Anyways, dozens PRs per day is draining energy. Don’t want to review anything anymore.

wrote:

I have no clue why the server hangs. I'm truly sorry about the downtimes. Any help would be appreciated.

wrote:

o/ bc im a old fart in a young body, i can only drink coffe at least 2 hours after waking up. So no until 10am

wrote:

Hey hola, aquí un hispanohablante que ingresó por curiosidad. No creo pasarme mucho por aquí pero igual aquí andamos por ahora.

Hey, hi! This is a Spanish speaker who joined out of curiosity. I don't think I'll be around much, but I'm here for now.

wrote:

Fair play to Apple and team, 5 apps in for review, 4 passed through each within an hour. 'Appy days

wrote:

What's happening

wrote:

me

wrote:

was shocked at how antiquated making international payments are in 2026!! used transfermate - no tracking, no visibility. just transfer and pray my money reaches 🙏

continued:

should have picked another payment method… silly me

wrote:

Hello, textlog!