Stephen Hardy

[Andrew Côté @Andercot]

It's pretty straight forward - OpenAI spontaneously emits @sama and @gdb, producing a short-lived @miramurati that converts to an @eshear via @bradlightcap decay.

@sama and @gdb undergo @satyanadella-catalyzed fusion with @MSFTResearch in a temporary bound state.

Since @eshear is a gluon he can mediate color-charge exchanges among board members: the @hlntnr exchange with @LHSummers preserves DC charge, and @TashaMcCauley exchange with @btaylor preserves SV-charge.

@adamdangelo is his own antiparticle and so remains in place, but traveling backwards in time.

The first half now plays out in reverse-time: the @MSFTResearch superposition collapses, @bradlightcap is re-absorbed along with @sama and @gdb into a new stable bound state.

The slight mass difference between exchanged board members adds 501(c) KeV of energy, causing net acceleration of the OpenAI ensemble.

[Stephen Hardy @proximation]

@iamtrask I thought this was interesting, regarding hierarchical representation- https://t.co/zBqHb1yyy1

[Zama @zama]

While OpenAI CEO @sama is speaking to the US senate about the risks of AI. reRead how we, at @zama are planning to solve the privacy issues of Large Language Models like ChatGPT with Fully Homomorphic Encryption
https://t.co/Zi8PTCtavU
#FHE #Privacy #AI

[Stephen Hardy @proximation]

@iamtrask https://t.co/ky9D0gKnx5 is probably the closest - Wikipedia is just too small to train a capable llm all by itself - but add arxiv and others…

[Stephen Hardy @proximation]

@ylecun This paper from January suggests that llms + memory are Turing complete… do you disagree? https://t.co/KCly015Kz5

[Stephen Hardy @proximation]

@hardbyte_nz This is super cool - cheap way of going from gpt to chatGPT : https://t.co/zfHnFk696K

[mjm @matthewjmclean]

One of the many passages of Hilary Mantel that will never leave. ‘All houses are haunted.’ RIP.

[MIT CSAIL @MIT_CSAIL]

RT if you were part of that tiny 1993 bar.

(via @markrendle)

[Eric Horvitz @erichorvitz]

On readying for new forms of #AI-powered manipulation & disinformation—and directions on mitigation. Forthcoming Blue Sky Ideas track article https://t.co/tDUBLpVLdv @scsp_ai @compcomcon @PartnershipAI @BBCWorld @nytimes @CBSNews @C2PA_org @truepic @senrobportman @SenGaryPeters

[George Nott @georgepnott]

21st Century Schizoid Spam

h/t @IridiumNYC

[Davis Blalock @davisblalock]

"Understanding Scaling Laws for Recommendation Models"

For two years, the AI world has had this glorious period of believing that big tech companies just need more compute to make their models better, not more user data.

That period is ending. Here's what happened: [1/14]

[Stephen Hardy @proximation]

There is such a gap between the pointlessness of existence and the pointlessness of non-existence. That gap is where all of life happens.

[Stephen Hardy @proximation]

@stilgherrian I proffer the Blues Brothers, with special emphasis on the sound track during the car chases.

[Zibo Chen @ZiboChen]

Can proteins perform neural network computation? Together with James Linton, @RonZhu2015, and @ElowitzLab, we had a lot of fun creating the "perceptein" -- a protein circuit that performs winner-take-all neural network computation in mammalian cells: https://t.co/uhFJoYGbJF (1/4)

[Yves-A. de Montjoye @yvesalexandre]

Scientists 👩‍🔬👨‍🔬: simply removing identifiers does not prevent re-identification. However, if you use modern privacy engineering techniques you can use data while properly protecting it.

UK gov: This sounds like an "impossibly high standard". Hold my beer 🍺

[Stephen Hardy @proximation]

@mkskeller Awesome work Marcel - under an hour!

[Marcel Keller @mkskeller]

A virtual workshop on privacy-preserving machine learning will take place after this year's @NeurIPSConf. You can now submit an abstract: https://t.co/Ba1CBeA8WV

[Cape.ai @cape_ai_]

Today, we're proud to announce the launch of our open-source platform and $5 million in seed funding! It’s our mission to make privacy-enhancing technology fast and easy to change the world for the better.

Thank you to @TechCrunch for the article: https://t.co/fGF28yCCdO

[IAG @IAGAust]

We’re working with Ambiata to deploy Atmosphere, a closed-loop #AI and machine learning tool that’s helping us drive better customer outcomes through personalisation
https://t.co/OqdapFnqA2

[Stephen Hardy @proximation]

@SciBry I am so sad he passed, but inspired that he pressed on over these 20 plus years and did something great. And sad again that he won’t do more. Vale JP.