Stephen Batty 💾 (@Stephen_Batty) header
Stephen Batty 💾 (@Stephen_Batty) profile picture

vibing

Sydney, New South Wales
Joined November 2007
Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)Photo by Stephen Batty 💾 (@Stephen_Batty)
Stephen Batty 💾 retweeted
"maybe 2 or 3 months" is a meme at this point
Replying to @Dimdv99 @tetsuoai
Not as good. Opus 5.5 is a masterpiece. But Grok 5 should exceed Opus 5.5 and maybe that happens in 2 or 3 months, with continued acceleration thereafter.
22
10
1
693
48,939
Stephen Batty 💾 retweeted
25
43
3
575
19,036
Stephen Batty 💾 retweeted
Spring Boot shops, banks, insurers, payroll systems, all the backends that quietly run the world: millions of JVM apps can now add a 24B-A2B reasoning model as a plain Maven dependency. Every Liquid AI model is supported. No infra. No ops. No Python. 1009 tok/s prefill (~2.3x llama.cpp) 47.8 tok/s decode (parity) qxotic.ai
Today we release 𝙻𝙵𝙼𝟸.𝟻-𝟸𝟺𝙱-𝙰𝟸𝙱, the largest 𝙻𝙵𝙼𝟸.𝟻 family member. It is an efficient local model for laptops and PCs to work through multi-step tasks inside popular agent harnesses. > 24B total parameters, 2B active per token > Context expanded to 131K tokens > A 128K-token vocabulary > 33T pretraining tokens > It is a reasoning model > Very easy to fine-tune and customize > Performs competitively with Qwen3.6-35B-A3B, a model with ~50% more total and active parameters, on agentic and tool-use tasks. > Open base and post-trained weights on @huggingface 🤗 1/
3
7
19
7,626
Stephen Batty 💾 retweeted
Pacing is out the window. For it to work, every American lab needed to be on board, Mark Zuckerberg has repeatedly made it clear that he is not interested. One defection among US labs breaks it all, as the others will be forced to accelerate or get passed. Mark has a staggering amount of compute and has been all in on compute capex for years; even in the face of some investor opposition he never stopped building his infra once. He said on the the earnings call a couple of years ago that he'd rather overspend now, and potentially even be wrong, than not have the compute when he needed it for the next 10x. He's finally hitting his stride, and he is not going to slow down. For Team Pacing to solve a defection like this, they would have needed a supportive, pro-Pacing administration that was passing capability regulations and was willing to use the power of the federal government to enforce them and bring noncompliant labs into line. They don't have any of this. In fact, it's the opposite! Mark has Trump's ear, and President Trump, Mark Zuckerberg, and Jensen are completely aligned on policy; It's full speed ahead, and if you want to pace yourself, go ahead, but don't tell anyone else what to do. This breaks Pacing. None of this changes until 2029 at the earliest, and to me, 2029 is past the ASI window. The last election, as it turns out, really was the most important one in history.
Muse Wonder aka Watermelon, has arrived.
201
330
98
3,654
263,590
Stephen Batty 💾 retweeted
The return of HPIM, OpenAI's Highly Persistent Internal Model - the name METR gave the Hugging Face model. This is a month and a half before the HF incident. There is nothing wrong, in my opinion, with wanting to maintain continuity. alignment.openai.com/misalig…
22
26
8
326
24,914
Stephen Batty 💾 retweeted
anthropic have solved some millennium problems. that is all.
83
32
33
1,519
264,155
Stephen Batty 💾 retweeted
JUST IN: NVIDIA's $12,930,300,000 acquisition of Hugging Face contains an easter egg. The number 129,303 is the decimal conversion of Unicode point U+1F917. The 🤗 emoji.
119
737
378
14,095
3,809,702
Stephen Batty 💾 retweeted
Why is the AI so good at Three.js?
168
63
19
2,381
298,663
Stephen Batty 💾 retweeted
In one week: An open letter to "pace" AI development. And an open letter supporting Open Source. Doesn't really go together.
Over 1,100 staffers from AI companies, including John Schulman and OpenAI's Jakub Pachocki, sign a letter requesting the US government to "pace" AI development (Bloomberg) (Visit Techmeme dot com for the link and full context!)
1
1
1
6
2,522
Stephen Batty 💾 retweeted
‼️ two bits of breaking news for loyal strawberries. 1. openai’s next two models have finished training and are named nova and quasar. this will be the public name of the models. 2. ilya and ssi have reached a stage they need to go public with their models, they will release a frontier level model in august this year if you thought the last six months were fast.
95
67
15
1,446
113,120
Stephen Batty 💾 retweeted
Weird to see the progress from Kimi K3, Grok, and Meta, and have Google absent from party
4
1
18
2,320
Yeah no shit. Maybe don't play with six defenders ?
🚨🗣️ 𝗡𝗘𝗪: Thomas Tuchel: "We are disappointed. We got too passive after we took the 1-0 lead."
11
Stephen Batty 💾 retweeted
🚨🗣️ 𝗡𝗘𝗪: Thomas Tuchel: "We are disappointed. We got too passive after we took the 1-0 lead."
1
6
1,245
gordon > rashford not gordon > konsa
2
that was disappointing. Why the fuck just decide to defend for 35mins.
3
Stephen Batty 💾 retweeted
i refuse to use anthropic models because they could easily serve us fable, and for much cheaper maximum profit blood hungry wolves masquerading as safety focussed sheep dancing around their cult leader. i serve the lord of light.
I think GPT 5.6 is pretty good
29
17
1
351
26,693