Mike Smith

AI researcher who likes self-supervised, unsupervised and gen models.

Was at @I_Hert_Space and @AspiaSpace, now at @astro_ai_cfa and @universe_tbd.

[Mike Smith @smith42mike]

Check out 100TB of machine learning ready multimodal astronomy data 🔥🦾

Excited to see what FOSS machine learning world does with this 🔭

[UniverseTBD @universe_tbd]

Excited to announce the Multimodal Universe (MMU): a huge 100TB dataset bringing together the largest ML-focused collection of astronomical observations ever assembled to accelerate open AI and astronomy research

https://t.co/UVbgdguXek

Think ImageNet, but for space 🧵

[Nous Research @NousResearch]

The strongest models are gated and access is granted only to a select few.

Hermes Agent now exposes MoA presets as virtual models, giving you capabilities beyond the publicly available frontier: 8% higher than Opus 4.8 and 11% higher than GPT 5.5 on our upcoming benchmark.

[Maciej Pióro 𐙼 @maciejpioro]

@tunguz That's bc you need to be a little bit stupid to go all in on scaling LLMs. If you're too smart you get oneshotted by some useless off-ramp.

[Susan Zhang @suchenzang]

this is what open-source looks like

[Ronan @Ronanchamberss]

Starmer probably joining Anthropic

[Nous Research @NousResearch]

Hermes Agent has reached 200,000 GitHub stars

Thank you to our contributors, supporters, users, and agents!

[Stella Biderman @BlancheMinerva]

@tszzl @basedjensen What are you talking about? They didn’t share shit about it. Like, genuinely zero technical detail about the “principles of scaling RL over CoT” I was at a talk by Noam shortly after it launched and he answered every question about that with “I can’t say”

https://t.co/dG6HUlq11c

[Mike Smith @smith42mike]

lmao "COMMON NAME PLACEHOLDER" 🌝

[John F. Wu @jwuphysics]

Genuinely surprised lol

https://t.co/WX8SMWu49W

[gelo @big_gelatin]

these deepmind resignations are getting out of hand

[Jeremy Howard @jeremyphoward]

Wow.

@Zai_org GLM 5.2 is a marvel! It is *at least* as good as Opus 4.8 and GPT 5.5. It's super fast, inexpensive, and not too verbose.

It responds with nuance and judgement, & handles long context VERY well.

I've never experienced an open weights model like this before.

[Chubby♨️ @kimmonismus]

Who needs Fable 5 when you got Le Chaton Fat by Mistral

[Mike Smith @smith42mike]

running out of room for the entire southern sky 😭

https://t.co/l8RDznYGXZ

[Mike Smith @smith42mike]

@BlancheMinerva @DanielCHTan97 i wonder what the oldest citation in ML is currently, we managed to squeeze plato into one of our papers

[Sara Hooker @sarahookr]

This technology was built of the shoulders of many.

Many will carry it forward.

Care deeply. But don’t dictate.

[Sriram Krishnan @sriramk]

@kzitouni1 I refuse to believe a small set of people however well meaning are safer than broad academic scrutiny.

[Mike Smith @smith42mike]

we predicted anthropic's fable's restrictions on AI and biotech research way back in 2022 https://t.co/BFm4rYujUb

the solution is the same: collective investment in open models and open datasets released under a "copyleft" licence for AI

[Mike Smith @smith42mike]

we predicted anthropic's fable's restrictions on AI and biotech research way back in 2022 https://t.co/BFm4rYujUb

the solution is the same: collective investment in open models and open datasets released under a "copyleft" licence for AI

[Nathan Lambert @natolambert]

I got a good nights sleep and I’m still just as angry about Anthropic’s choices.

I enjoy working in AI so much and to have my access to the cutting edge models for my work rugpulled in an under the table fashion is appalling.

I expected to be restricted eventually, but not now, and to be told it directly.

[Mike Smith @smith42mike]

See "The Case for Contextual Copyleft: Licensing Open Source Training Data and Generative AI" https://t.co/at5qB4nB8s we need a GPL for the AI era.

[Mark Saroufim @marksaroufim]

June 9th Researcher Reciprocity License

"if you train on it, you let us generate - reverse terms of use void"

Status quo
1. We teach frontier devs with ICLR/NeurIPS papers, OSS Github contributions
2. They use it to make frontier models
3. Then ban us from exploring our ideas

We need a new license, original thinkers can't be an underclass to a tyrannical researcher fiefdom

[Zy @ZyMazza]

This is the playbook. It works because life is long and memories are short

[Daniel Auras @rasdani_]

this is the biggest wake-up call to protect and nourish open source AI

if you don't build out sovereign and independent models+infra closed labs will patronize you to an insulting degree