Stephen Mayhew (@mayhewsw) header
Stephen Mayhew (@mayhewsw) profile picture

๐˜ฑ๐˜ญ๐˜ฆ๐˜ข๐˜ด๐˜ฆ ๐˜ค๐˜ฐ๐˜ฏ๐˜ด๐˜ช๐˜ฅ๐˜ฆ๐˜ณ ๐˜ต๐˜ฉ๐˜ฆ ๐˜ฆ๐˜ฏ๐˜ท๐˜ช๐˜ณ๐˜ฐ๐˜ฏ๐˜ฎ๐˜ฆ๐˜ฏ๐˜ต ๐˜ฃ๐˜ฆ๐˜ง๐˜ฐ๐˜ณ๐˜ฆ ๐˜ฑ๐˜ณ๐˜ช๐˜ฏ๐˜ต๐˜ช๐˜ฏ๐˜จ ๐˜ต๐˜ฉ๐˜ฆ๐˜ด๐˜ฆ ๐˜ต๐˜ธ๐˜ฆ๐˜ฆ๐˜ต๐˜ด

Joined December 2011
Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)Photo by Stephen Mayhew (@mayhewsw)
Iโ€™m excited to announce an ambitious project: โœจUniversal NER โœจ. Multilingual NLP is missing a gold standard multilingual NER dataset, and this project aims to fill that gap. This is a community effort: I need your help! See: universalner.org Or read more in the ๐Ÿงต
12
56
4
292
Stephen Mayhew retweeted
the fact that grass can become a sweater through the intermediary of a sheep ought to astonish us more than it does.
the average person doesn't appreciate wool enough
110
1,628
64
24,628
269,087
the biggest promise of AI is allowing humans to give instruction to machines -- a natural interface to a useful tool. I like to call this "automation in the Information Age" These products do the inverse: giving machines an interface to interact with humans.
After Google acquired my AR startup, I started thinking about the next interface, this time for AI. Text and voice work fine, just like DOS did before the Mac. But nothingโ€™s more intuitive than face to face. After two years of research, today weโ€™re launching Atmee v2: a real-time avatar model that makes live video agents feel truly life-like. โ€ข Any character from 1 image + 10s of audio โ€ข Contextual expressions, gestures, even interacting with the environment โ€ข Live video calls, voice only, or video generation At its core, Atmee v2 is a video foundation model, it can render anything, but itโ€™s built to excel at characters, which is what video models like Seedance 2.5 and Wan 3.0 are mostly used for. In comparison: Atmee v2 โ€ข Up to 50x faster โ€ข 10-100x cheaper Build tutors, sales coaches, support agents, companions, game characters, even whole living worlds. The interface after text and voice starts today.
1
141
and, look, I know it's all going to be different in "3-5 years" etc. etc. but it's also just so unconvincing. If anyone asked me to interact with this kind of thing, I would take it as an insult.
15
> listened to a track (my own recording) on Soundcloud > a new track comes on > pleasant, soft, piano jazz > I like this, what is it? > it's AI > RIP music recommendations, was fun while it lasted
101
Stephen Mayhew retweeted
i've reproduced the results from the MIT blog post today on weight-space programming. but one additional result i have to add: this is a new, step-change emerging capability that's appeared in just the latest set of models.
218
803
268
7,819
600,833
Stephen Mayhew retweeted
Asking my AI teacher how to spell seventeen
3,017
13,027
4,437
208,681
28,771,615
the class of things we (i mean i) consider as "meh, of course an AI can do *that* " these days is kinda insane tbh.
2
1
1
55
7,673
there's no way this can be real
We asked GPT-6 Astra to edit Qwen3.5-0.8B weights to improve math. It directly changed only 0.016% of weights, fixing 0.8B's reasoning loops & improving math up to 6x and even instruction following 3x, without gradients or training. We call this Weight-Space Programming: frontier LLMs can now natively read/write neural network weights! ๐Ÿงต Blog: rikiyat.github.io/blog/weighโ€ฆ
Readers added context they thought people might want to know
LLM was given weights explicitly as text in prompt and used code/forward passes (per authors' logs) to ablate Qwen looping direction, a known thinking-mode issue per model card. Primarily fixed repetition in that mode, not general 'natively' improving weights/math. rikiyat.github.io/blog/weight-spโ€ฆ github.com/RikiyaT/lm_undโ€ฆ huggingface.co/Qwen/Qwen3.5-0โ€ฆ huggingface.co/RikiyaT/Qwen3.โ€ฆ
100
Stephen Mayhew retweeted
I'm on the industry job market this year! I'm a PhD student at CMU LTI. My work spans post-training, evals, alignment and multimodality! Please reach out (DM) if you are hiring and think I'd be a good fit, I'd love to chat :) Website: simran-khanuja.github.io/ CV: simran-khanuja.github.io/assโ€ฆ
11
69
3
484
41,116
Stephen Mayhew retweeted
I love that scene in Good Will Hunting where Ben Affleck is talking to Matt Damon's titular character Will Hunting and he's like โ€œyou've got a gift man and you've gotta get outta this town and make something of yourselfโ€ and the gift in question is being really good at math and the town he needs to get out of is Boston, Massachusetts.
440
1,885
181
74,876
2,718,085
Stephen Mayhew retweeted
There's an old joke about a scientist who tells God the creation of the world wasn't any great shakes. He could do just as well himself. Whereupon he picks up a handful of dirt. "Not so fast," God says. "Get your own dirt." Claude, get your own dirt. /End
1
13
1
67
3,795
is there a more poignant picture of the โ€œAI revolutionโ€?
New record for Stack Overflow: <1000 questions posted last month
1
167
my new measure of techno-modernity is: how many battery-powered devices do I interact with each day? yesterday: at least 9 (scale, phone, laptop, smartwatch, air-pods, over-ear headphones, metronome, lawnmower, iPad)
2
1
2
271
allow me to share this original organic poem: they said modern life would be freeing, enlarging but Iโ€™ve got 99 devices and they all need charging
2
30
Stephen Mayhew retweeted
Oh, my, @nytimes .... again?!? ๐Ÿ™ˆ๐Ÿ™ˆ๐Ÿ™ˆ
How does this nonsense make it into the paper of record?
1
1
17
1,834
I just executed on my โ€œchange the profile pic every 10 yearsโ€ plan. see you again in 2036
6
235
Stephen Mayhew retweeted
Disconnecticut
44
272
27
5,046
105,244
Stephen Mayhew retweeted
I just found out the GPT in ChatGPT stands for Gomputer Pomputer Tomputer
98
2,873
124
35,782
845,190
wow, very glad Iโ€™m missing this
100% of the presentations I've seen so far at EMNLP have used AI generated slides. I'm not opposed to using AI tools, but most of these slides have been very low quality...
2
7
780
Stephen Mayhew retweeted
My historical methods professor in grad school told us he was learning to code because "pretty soon," he said, "not knowing how to code will be like not knowing how to read." Turns out he was right Because now it seems nobody is going to be doing either of those things
4
32
6
423
8,125