I'm launching a newsletter, called @insight. I want to try to create a space for thinking deeply about complex, thorny but important puzzles, along with a community. Something between public writing and social media: some of both but neither. Sign up here:

British-Egyptian writer @alaa has been officially pardoned by Egyptian president @AlsisiOfficial following an appeal by Egypt's National Council 4 HR (long been been silent over systematic HR abuses). Looking forward to seeing Alaa united with his family in the #UK

Can’t beat them? Connect your broker to trade like them.

#Egypt: CPJ welcomes Egyptian President Abdel Fattah al-Sisi’s decision to consider a pardon for Egyptian-British writer Alaa Abd el-Fattah. Earlier this year, CPJ, along with 50 prominent global figures, called on President Sisi to grant this pardon. It is time for Alaa to be

#Egypt: Egyptian-British writer Alaa Abd el-Fattah’s resumed hunger strike is a tragic reminder of his continued arbitrary detention. Egyptian authorities must release him immediately and stop using his imprisonment as political leverage, especially after removing him from the

All this could happen without any reference to religiosity or invoking a mind/body duality. Ironically, asserting computational functionalism as an argument for moral standing or its equivalence seems to require a specific (and new as of 20th century) type of mind/body dualism.

What if there were no more mass shootings... What if there were a cure...

So even if most people accepted most everything asserted by computational functionalism, my guess is most still would categorically reject that as an argument for moral or equivalence or standing simply because people already have a strong sense of what the equivalence layer is.

I’d guess the “substrate” approach would extend to animals for many people. The “substrate” with shared mortality and fragility vs high information processing a la Shannon? “Either your pet dog gets tortured or killed or we unplug this machine smarter than your dog for a week”

Here, too, depends on whether “mind” is meant to imply human moral standing equivalence or just surpassing some high threshold of complex information processing. One can argue the former is substrate-dependent regardless of whatever a machine system measures under the latter.

“A simulation of a hurricane isn’t a hurricane” is indeed only an assertion, but it’s not what Searle says. Searle argues a simulation of a hurricane will not leave us drenched. Very different. For AI, would have claim humans are reducible to information, in the Shannon sense.

Hydrate. Hustle. GO! CELSIUS HYDRATION - The ultimate hydration for every move. CELSIUS. LIVE. FIT. GO!

Obligatory RT of any mention of Herb Simon.

Just saying “crypto” means cryptocoins nowadays. So I named: zero-knowledge proofs, distributed ledgers, etc. And secure enclaves and hardware tokens are not just crypto. But yes, THIS IS something for cryptographic methods! The alternative is uber-surveillance by authorities.

No, it doesn’t because other things we used back then are also gone. The printing press may be a great invention, but it also unleashed centuries of chaos that killed more people than all the other wars since. Easier is a source for major transition tumult. Transition is hard.

Maybe in the long run, fine. But the transition? Such chaos can fuel nightmare level authoritarianism. We need to switch to authentication and verification at source, rather than post-hoc. Distributed ledgers, zero-knowledge proofs, secure enclaves etc. are all great for this!

Herbert Simon noted in 1971: info glut makes attention scarce, a historical inversion. Technology turns hard into easy, rare into common, scarce into abundant. But we often use the former as a safeguard. That friction is load-bearing. Tumult ensues. 🎁

AI video is almost there: can’t tell fake from real. Losing our last method of post-hoc verification is no joke. When technology makes something easy, it can also knock down load-bearing obstacles where the difficulty was a safeguard. That’s what causes the transition chaos.🧵

Win LA, win the world. 🌎 @flyarcher's strategic acquisition of Hawthorne Airport will give us an incredible asset in one of the world’s most important eVTOL markets. We’ll look to transform the airport into our hub for Archer’s Los Angeles air taxi network as we ready for

And today’s AGI brouhaha deploys thinking and language developed by the symbolic school, partly for their war against their connectionist nemeses, but now serves as obfuscation on behalf of a (spectacularly successful, but still) connectionist thrust of AI. I’d watch the movie.

“What did First-Name-in-AI see?” Something very interesting maybe, but I would put my money on fame, fortune, glory, fantasy, outsized early stage funding, a story to help distract journalists and investors, etc. Story as old as time, as the saying goes.

Also no modern science during the Pleistocene! Humans are embodied beings with specific cognition, culture, education etc. We live and die in a physical world as evolved beings who inherit knowledge. For LLMs, AGI is a bit like saying cars are horseless carriages. Mismatch.

“But humans don’t reason strictly or well either.” No kidding. But we also aren’t just text. Reasoning and logic are specific skills, and gradient descent generating plausible versions of it after being trained on total output of humanity is impressive and useful but distinct.

Please click onto the image to hear oldies and classic rock

Generative AI is a plausibility engine. Sure, amazing. But see paper for umpteenth reason why understanding might soar if we ban the term AGI as a claim or benchmark and looked at specifics instead. Currently, “AGI” is a tool of obfuscation directed investors and the public.

Fair enough: It’s pre-print and I didn’t mean to pick one single paper (where they do put in limitations). It’s more the (very) general mindset that’s baffling: new major tech does not just replace its older version (or its previous human/animal doer). It’s waaay more turbulent.