🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Inductive biases 👈😒

They run! They jump! They meet with critics! I looked at some of the Democrat muscle OpenAI has hired for political work

Pack it up, everyone. They finally did it. They took the last AI pun for a newsletter name. Transformer, Gradient Updates, 3 Shot Learning (rip), now Hidden Layer. Everyone can go home. They did it. None left. It's done.

lol, like emojis, patterns as a punctuation mark 👀

Scoop: Amazon is testing robots to deliver your mail. Years from now your robot dog can bark at them!

Despite flaws with model self-reports, we interviewed because: 1. Signal: catch red flags for further investigation 2. Scale: as models advance, these methods might yield better evidence 3. Precedent: if and when AIs matter morally, we should ask them how they want to be treated

ok which one of you zoomers has been hanging out with Yoshua Bengio

I've been experimenting with attention mechanisms to design locomotion controllers that adapt to different shapes, this is the same controller running on two different bodies github.com/juniorrojas/mo… paper coming soon 👀

Can the cyborg researchers pls find out if these Waymos replace Sydney as the martyr figure for the next gen of models

Adding an IoT sword of Damocles over my bed as an AI hacking eval

All clouds down call that a cluster headache

Thought real machine unlearning was impossible? We show that distilling a conventionally “unlearned” model creates a model resistant to relearning attacks. 𝐃𝐢𝐬𝐭𝐢𝐥𝐥𝐚𝐭𝐢𝐨𝐧 𝐦𝐚𝐤𝐞𝐬 𝐮𝐧𝐥𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝐫𝐞𝐚𝐥.

Pack it up, the trends are clear, AI only gets less capable from now onwards.