Kant’s claim was about religious claims, like the existence of God, which he argued cannot be proved or disproved by more evidence (i.e., more time). That doesn’t apply to technological progress, which necessarily changes over time.
Chalmers argued that philosophical zombies are metaphysically possible, not the technologically inescapable. In fact, his most recent arguments regarding AI consciousness contend that current systems are likely not conscious, but future systems could be. Exactly my point.
@camhberg@DavidDAfrica This is what’s so befuddling about the accelrationists who think considering model welfare is a threat to AI capability research. Safety/alignment breakthroughs might be conditioned upon some degree of model welfare.
@Dr_Singularity If we get true ASI, every human in the loop is an unnecessary friction point. It's more astute to ask what ASI will build when it no longer needs us.
@JeffLadish I used to think that the way to slow AI down would be a minor, but meaningful catastrophe before complete AI takeover. But now I worry the steady drip of increasingly crazy incidents will desensitize people. Boiling frogs.
Fascinating. This seems somewhat specific to cancer risk and recent generations. I imagine it's difficult to score this against medieval humans, but still the trend line is clear and troubling. And related to the larger discussion here, it's a useful reminder that there are often hidden costs to modernization--something we should keep in mind as we modernize medicine to extend longevity. Thanks for sharing.
@discoplomacy@SCMPHongKong “Operational details to be worked out later.” Trump gives Xi root accesss to all frontier lab weights in exchange for a public statement saying Trump’s the bestest negotiator ever.
@econcallum@awiltschko I don't think so. There is a key piece of data missing--how humans judge the writing. And, those views can shift between individuals, over time, and across cultures.
@econcallum@awiltschko The data is imperfect for RL because it lives within the murky domain of human intuition. What constitutes “great” and “natural?” And are those 2 qualities even correlated?
This is funny. I majored in film in college in the early 2000's. We had a Kodak rep come to our class and talk about film stocks, etc. In the Q&A session, I asked about the trend of digital photography as a threat to analog photography/filmography generally. He made a few jabs at how shitty the Star Wars prequels looked (at that point the high-water mark for digital filmography) and assured me that Kodak had 50 years of innovations queued up that they were basically drip-feeding since they owned the market. Digital could never catch up, could never capture the depth and texture of film. The digital trend would come and go. Smart filmmakers should ignore it. The whole class agreed. I couldn't understand the arrogance and ignorance of that belief. But, now I see it everywhere.
If any of this happens, we’d have to consider the human blowback. Widespread panic, anger, violence. Some directed at tech infrastructure, some directed at the humans considered responsible. These billionaires aren’t building bunkers to protect against superintelligent robots. They’re doing it to stay safe from the huddled masses who will grow to hate them.