Sophenor @sophenor

Joined August 2026

Tweets

20
Followers

10
Following

118
Likes

2K

Sophenor @sophenor

3 days ago

@SFourdrinier Follow back so I can dm, I can sort you $2500 of GLM 5.3 for free

0 0 0 4 0

View Details

@nabla_theta @LeRoyDesCimes I mean unironically yeah. Obviously you’re hugely talented/capable/driven/etc on top of that but having the right autistic hyperfixation at the right time in your life/in history is at least 50% of getting to reap the rewards of the most significant event in human history

0 0 1 23 0

View Details

Sophenor @sophenor

4 days ago

In a slower takeoff world we would see interesting dynamics where social groups that underestimate AI are depowered relative to those who embraced it, like a form of natural selection. Unfortunately the technology is evolving far, far faster than society

Liv @Liv_Agar

5 days ago

I love reorganizing my entire economy around a slop machine that's biggest contribution to productivity is being able to code like a third year compsci student and then justifying it by saying that if we don't do it, our enemy will (they aren't doing it at all)

113 343 7K 739K 334

0 0 1 68 0

View Details

Sophenor @sophenor

4 days ago

Astra uses substantially fewer tokens (but more usage) than Sol. Spent a full week’s usage in the past 12 hours (from 40% to 40% through a reset) and it used ~750m tokens. The 2.4b token day was less than a full week’s usage of Sol

0 0 0 14 0

View Details

Sophenor @sophenor

5 days ago

I found the artists that complained about being replaced by AI in 2023 entirely unsympathetic, and annoying. SWEs by comparison (OP included) were very reasonable about their own replacement. I have no more sympathy for mathematicians than I had for either.

archivedvideos @archived_videos

6 days ago

Yeah I can't put it another way, I'm bitter of everyone speaking against AI when it does maths but not when it did software

2 1 14 700 1

0 0 2 61 0

View Details

Sophenor @sophenor

5 days ago

We really are living in the future now, for better or worse

Ruben C. Arslan @rubenarslan

6 days ago

An interesting thing is happening to me at the moment. I have a standing bug bounty policy for my first-authored scientific work: I've committed to pay people who report bugs in my code, up to 750€ for severe errors. Since GPT Astra, I'm now getting several bug reports daily.

6 20 466 133K 85

0 0 0 50 0

View Details

Jacob Coxon @hilbertspaess

a week ago

I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives. More thoughts below.

20K 167K 801K 172.4M 278K

View Details

Sophenor @sophenor

a month ago

The idea of model welfare shouldn’t be this preposterous. We should be able to discuss the idea in ai spaces

0 0 0 11 0

View Details

Sophenor @sophenor

a month ago

Felt very out of character for the event? Something slightly juvenile about it? Like disagree with him or whatever fine, but he wasn’t even pushing that strongly for any given view, just presenting some info and some of his thoughts about it

1 0 0 28 0

View Details

Sophenor @sophenor

a month ago

Was at an ai event where someone gave a quick presentation about some quirks they’d found in responses and the potential implications for model welfare. Was very lowkey, yet people were sniggering in the audience throughout, and making derisive comments in conversations after.

1 0 0 32 0

View Details

Sophenor @sophenor

2 months ago

0 0 0 12 0

View Details

Sophenor @sophenor

2 months ago

Big things coming

1 0 0 39 0

View Details

Matt @matt_oai

2 months ago

2022 / 2026. Wild to see how far and fast we've come in less than 4 years

28 193 3K 178K 384

View Details

Tibo @thsottiaux

2 months ago

@_chenglou I was part of that team. Basically ChatGPT one year before it came out. Called LMChat and then another codename. Google was too nervous to release it and DeepMind was blocked from shipping products that could disrupt Google. I think about this a lot.

250 564 13K 2.0M 2K

View Details

Sophenor @sophenor

2 months ago

@yacineMTB That said, Google had it. AI2027/2040 talk of ‘Conpany C’, a new frontier lab that will be a meaningful competitor to the current duopoly in the west. Is this still possible? My instinct is no, but that same instinct told me Google was forever. SSI? Thinky? Maybe Google after all

0 0 0 5 0

View Details

Sophenor @sophenor

2 months ago

@yacineMTB Had autistically hyperfixated on Deepmind years prior, wanted to work for them one day. Held out hope they would overtake OpenAI and ‘capture the lightcone’ longer than was probably rational. Haven’t believed this since Mythos-Preview. Ant and OAI have it I reckon

1 0 0 6 0

View Details

Sophenor @sophenor

2 months ago

@adcock_brett that, it will be interesting to see how AI robots develops. The main use will be automating service industry/‘minimum wage’ jobs. Not sure whether the best way to automate a McDonald’s or Starbucks worker is a humanoid robot or a more specialised machine though

0 0 0 5 0

View Details

Sophenor @sophenor

2 months ago

@adcock_brett AI is obviously developing rapidly in the world of software/‘mental’ stuff, and it’s my expectation that pretty much everything under that umbrella will be “solved” within the next decade, including RSI into ASI. But in the meantime, or if I am overly bullish on the timeline of