"I resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic. Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives... AI could kill us all by the end of the decade." Jacob Coxon Weekly view points. Could AI poses an existential threat to humanity by the end of the decade Which team are you on? Team PRO (Agrees with Coxon: The Risk is Real & Immediate) or Team CON (Disagrees with Coxon: The Fear is Overblown or Misguided) We want YOUR voice for the Mindplex Spotlight! Record a 1 to 2-minute vertical video sharing your raw take. Are you Team Pro or Team Con? 🌟 What’s in it for featured creators? Mindplex Spotlight Feature: All debate videos will be featured on Mindplex Magazine and Mindplex Youtube. How to join: 1, Record your 1-2 min video (clear audio, portrait mode). 2, Submit your video docs.google.com/forms/d/1o1H…

Mindplex has a new look & feel, a new state of the art website based on Typescript, and a brand new artificial intelligence (AI) -powered newsroom. magazine.mindplex.ai/post/we…

Discover how lightweight graph models predict online purchases up to 180 times faster than heavy deep learning algorithms, outperforming complex neural networks with simple efficiency.

What if AI evolved by optimizing what code actually means, rather than just tweaking syntax? Discover how Semantic Genetic Programming revolutionizes the future of AI.

Reading Amodei’s post, the words that come to my mind are virtue signaling and regulatory capture. The result would be the opposite of what they say they want.

This post collects links to other interesting news that we have considered but not included in our regular news this week.

Artificial intelligence programs reviewed tens of thousands of human drug studies, found a pattern tied to success, and proposed a lung cancer design later taken up by a manufacturer.

MIT researchers place an ultrathin layer of PDMS between metal plates so voltage-driven motion stores a short history of inputs and then fires like a biological cell.

An AI tool rebuilds a published study from its text, figures and data so the resulting agent can explain the work, reuse its methods and compare notes with other papers.

The future of artificial intelligence is no longer being decided only in laboratories of algorithmic design. It is also being decided in wafer fabs, high-voltage grids and server rooms.

Researchers propose a charge-neutral helium-3 oscillator that may resist electrical noise, with far lower predicted error rates than superconducting qubits.

TranscriptFormer was trained on 112 million cells and can compare types across species, including cells the model has never seen.

Collective changes in tiny magnetic textures produce stable artificial synapses at room temperature and could cut the energy used by neural-network operations.

Organic sheets printed from room-temperature inks switch in 26 nanoseconds, recognize handwritten digits almost as well as software, and wash away after use.

My guess is that the remaining Millennium Prize Problems, and then large stretches of physics and the other formal sciences, will go the same way. But so what?

Researchers searched hundreds of millions of protein shapes and found TM184C, an inner-cell protein that helps neighboring cells exchange materials through thin bridges.

A Nature study in mice finds that a tiny class of long-range inhibitory cells in the cortex can slow and synchronize brain activity and increase sleep.

This post collects links to other interesting news that we have considered but not included in our regular news this week.