Todd L Corenson

Looking for the next little thing that could become the next big thing.

[Todd L Corenson @ToddLC]

The inaugural @fv_summit really Leaned into banning AI hallucinations. #software #math #AI #Lean

[Todd L Corenson @ToddLC]

@torgeirlysen @littmath @SidGadgil I suspect even if they gave us a formally verified proof, we would be disappointed that humanity hadn't discovered it.

[Todd L Corenson @ToddLC]

@cenkuygur Next step: capture Qeshm and Larak. There's a new toll troll in town!

[Todd L Corenson @ToddLC]

@ylecun @DavidSacks How much of the canceled research was DEI-oriented or awarded?

[Todd L Corenson @ToddLC]

@nicksortor Enough already. Time for the U.S. to take over Qeshm and Larak. Then the U.S. will control the toll booth and the whole strait. There's a new troll in town!!

[Todd L Corenson @ToddLC]

@_quasi @GaryMarcus Well for one thing, it’s theft.

[Todd L Corenson @ToddLC]

@annagrad78 If consciousness has no power as viewed from the outside, a computable AI can in principle do everything that a conscious being can. Alternatively, if this is not the case, we will need to find a new substrate for AI if we want to achieve AGI.

[Todd L Corenson @ToddLC]

@annagrad78 Sure. From an objective perspective, the interesting question is what power does it have? What can it do, if anything, beyond what is computable? This should be answerable before we have a true understanding of how to create a subjective experience.

[Todd L Corenson @ToddLC]

@annagrad78 No, by definition if something is truly computable, it must follow the rules of a Turning machine. The only loophole is that the something could have epiphenomenal experience.

[Todd L Corenson @ToddLC]

@annagrad78 We know that brains are composed of neurons. But we don't fully understand the mechanism. Conscious "escape" could be the result of Penrose Orch OR or something else.

[Todd L Corenson @ToddLC]

@annagrad78 The distinction is that we know the underlying process for AI, but don't for humanity, and likely, life in general. We know that AI is a Turing machine. The only escape is if the mechanism is broken and not error correctable.

[Todd L Corenson @ToddLC]

@annagrad78 Also, what we know is that consciousness in humans can't be just computable as in AI. Otherwise it would merely be an epiphenomenon, which at least introspectively doesn't seem to be the case.

[Todd L Corenson @ToddLC]

@annagrad78 If you're arguing that everything in AI is not predetermined from the start, what you are saying in essence is that matrix multiplication is failing when a system is sufficiently complex. And that it's failing not randomly, but in such a manner as to manifest consciousness.

[Todd L Corenson @ToddLC]

@annagrad78 So the question is, is real consciousness orthogonal? And is there an eval that can unambiguously determine this?

[Todd L Corenson @ToddLC]

@annagrad78 No matter how fancy, current AI is classical computing. It’s running a program. So this mystical consciousness cannot be anything other than an epiphenomonym. A philosophical zombie would behave in precisely the same way.

[Todd L Corenson @ToddLC]

@mb_ghalibaf The obvious alternative is for the U.S. to win the war. This can be accomplished if the U.S. takes over the Persian Gulf including the Strait of Hormuz, specifically Qeshm and Kharg islands. Then only allied oil transits.

[Todd L Corenson @ToddLC]

@WitoldZiarno @spaceandtech_ Not if it throws one first.

[Todd L Corenson @ToddLC]

@araghchi Project Freedom is Project Boa Constrictor.

[Todd L Corenson @ToddLC]

@bulkeley_john @EricLDaugh So what's a fair escort fee for this service?

[Todd L Corenson @ToddLC]

@moonsandhues Chalmers -> AGI -> economy ?? That has possibilities.