Ol Ceru

[Miriam Franchina @barlafuess.bsky.social @Barlafuess]

Angesehene Anerkennung @RD100Awards für Innovation für Forschende @AttilaCangi & Lenz Fiedler @CASUSscience #Görlitz/@HZDR_Dresden
Im Video: wie #künstlicheIntelligenz mithilfe der Software #MALA für #Materialwissenschaften angewendet wird:/1
https://t.co/Txzwg3F0lk
@MDR_SN

[François Chollet @fchollet]

This paper has the right idea: use symbolic logic for discrete reasoning and lean on deep learning models for perception and common-sense intuition. https://t.co/9lP8eDZKkO

I expect to see a lot more progress along these lines in the coming months / years.

[Samuel "curry-howard fanboi" Ainsworth @SamuelAinsworth]

latent codes are isometric => zero-shot stitching of independently trained encoders/decoders... cool stuff!

one day we're gonna wake up and discover deep learning is secretly just some kernel methods in a trench coat

[Rohin Shah @rohinmshah]

Excited to release our work on Goal MisGeneralisation (GMG)! In GMG, a model trained with a correct reward function pursues undesirable goals. E.g. this agent competently gets negative rewards.

Blog: https://t.co/y9tAAbLWEp
Paper: https://t.co/lFADb7pCyB

https://t.co/bToRP17VIE

[Riley Goodside @goodside]

Exploiting GPT-3 prompts with malicious inputs that order the model to ignore its previous directions.

[Anthropic @AnthropicAI]

Neural networks often pack many unrelated concepts into a single neuron – a puzzling phenomenon known as 'polysemanticity' which makes interpretability much more challenging. In our latest work, we build toy models where the origins of polysemanticity can be fully understood.

[Joshua Achiam @jachiam0]

I feel very conflicted about the Stable Diffusion open source release.

[François Chollet @fchollet]

You can tell crypto is a great investment because of all the billboards and TV ads telling you to buy it (~$1B ad spend in 2021).

Those paying for the ads want to share with you the fabulous profits to be made in crypto. They definitely don't need you to be their exit liquidity.

[Corrado Monti @c0rrad0m0nti]

"While 66–80% Americans support [major climate change mitigation] policies, Americans estimate the prevalence to only be between 37–43% on average."

[François Chollet @fchollet]

"would this still matter in a world where money didn't exist" is a useful question to ask about most things

[Evgeny Morozov @evgenymorozov]

A very powerful anti-ITT cartoon from Allende-era Chile

[François Chollet @fchollet]

Over the past 10 years, AI has specialized into two domains.

"Applied AI" where the goal is to ship products with real-world impact.

"Make-believe AI" where the goal is to ship press releases to convince investors and the public that what you're doing is important.

[Peyman Milanfar @docmilanfar]

when everybody is chasing the state of the art

[François Chollet @fchollet]

I'm hearing disturbing rumors that a number of fabulously rich men are turning out to be very right-wing and are seeking to buy up media outlets in order to increase their political influence and shore up their views. This is unprecedented and no one could have predicted it.

[Richard Ngo @RichardMCNgo]

Asked a ML researcher friend what would convince her that large models aren't "just reflecting their data" - she said understanding novel similes and metaphors.

Here's one that GPT-3 basically nailed - but it's probably not fully novel. Other suggestions?

[Ol Ceru @il_ceru]

@matteosalvinimi Che guidi tu?

[Google DeepMind @GoogleDeepMind]

Introducing #AlphaCode: a system that can compete at average human level in competitive coding competitions like @codeforces. An exciting leap in AI problem-solving capabilities, combining many advances in machine learning!

Read more: https://t.co/yaXfMWtEfe 1/

[Evgeny Morozov @evgenymorozov]

Has anyone noticed that the one-trick-pony theory of "surveillance capitalism" can't offer any critique of "Web3"/crypto? At best, it would embrace them as a form of "humane" informational capitalism - the "creator economy" and all that - which is finally reasserting itself.

[François Chollet @fchollet]

I think I'm pretty close to understanding how music works. *Why* music works.

[Caroline Orr Bueno, Ph.D @RVAwonk]

omfg this conversation between a researcher and a QAnon-primed GPT-3 model (AI language generator). The researcher is asking the questions; GPT-3 is answering them.

https://t.co/vzWIVqUQzX