441–450 of 595 posts
Earlier quoted context omitted.
If you expand your timeline a little bit you'll reach a spot where the computers, equipment and comms necessary to create a computer network were highly centralized and only run by large corporations and institutions. How many times do we need to see this pattern for it to be obvious?
History does NOT NECESSARILY repeat itself.
Earlier quoted context omitted.
Source? My understanding was this came from within and had nothing to do with products but instead how employees treated one another.
Tim Ferriss interview with him. It came from within, maybe I didn't explain myself clearly.
Earlier quoted context omitted.
Almost certainly true but the idea that the world is made safer by Cyberdyne open sourcing lots of it's dangerous technology and probably spawning many more mini-Cyberdynes strikes me as extremely naive.
If the doomsday scenario is one AI going rogue because of misaligned goals, then having lots of AIs going rogue in various different ways seems indeed preferable, because the AIs will compete with each other and neutralize each other to some extent.
At least with a single AI, there's a chance that it will leave something for humans.
Earlier quoted context omitted.
What strong evidence of existential threat are you expecting to see before it's already too late to avoid catastrophe?
What strong evidence do you have that AGI (1) is possible in the foreseeable future, and (2) will be more a threat than any random human?
At some point the U.S. government should step up and provide some resources, and offer large datasets for open acess training and pretrained models. In the future, there will be many AI models for various purposes, closed source, guarded, with their moat being the costs to train a similar model on a dataset of similar breadth. This eliminates smaller orgs from ever being able to offer a competing model, and ruins inn…
Dunno how many of them are thinking "let's release this for EVERYONE"
Including our adversaries, many of whom are technologically behind us
The only way this kind of research could be done "in the open" is if it was funded by taxpayers' money. Except we have global corporations with elite fiscal layer teams, but no global government (giant lizzards are always disappointing) - so some global corporation was going to own it in the end, it was a matter of "time" and "US or Chinese". The time is now, he winner is US. Moving on. The next move is for (some) go…
> funded by taxpayers' money. In this industry, a contract with MS is as close to directly having a DOD contract as you can get. Remember when MS bought Skype, turned off end to end encryption and got rid of the P2P mechanisms, to eventually appear in the leaks about PRISM? Likely Bing integration of ChatGPT is the more innocuous use of the technology. Once they have gotten input from the AI experts at OpenAI on how…
I find it a little odd that Elon seems to take a swipe at OpenAI any opportunity he gets. If he cares so much about them not making money, maybe he should have put his twitter cash there instead? It's reassuring to me that the two people running policy work at the big AI "startups", Jack Clark (Anthropic) and Miles Brundage (OpenAI, who was hired by Jack iirc), are genuinely good humans. I've known Jack for 10 years…
"OpenAI was created as an open source (which is why I named it “Open” AI), non-profit company to serve as a counterweight to Google, but now it has become a closed source, maximum-profit company effectively controlled by Microsoft. Not what I intended at all." - Elon You can think what you want of Elon, but he is in the right here.
Earlier quoted context omitted.
Are there any companies like Docker? They feel like such a Black Swan to me; namely -- "billions" + very useful + pretty much "not evil " at all. I literally can't think of any others. I don't mean "literally evil" of course, I personally acknowledge the need to make money et al, but I mean it even seems like your most Stallman-esque types wouldn't have too much of a problem with Docker?
While I have no problem with Docker, it probably is worth noting that their entire product is based on a decade and a half of Google engineering invested into the Linux kernel (cgroups). Not that that demeans or devalues Docker, but it contextualizes its existence as an offshoot of a Google project that aimed to make it possible (and did, but only internally)
Earlier quoted context omitted.
The difference is that Tesla isn't claiming to be open about AI.
And resigning because he was working on a competitor was completely the right thing to do. It would have been a conflict of interest. How often does that type of integrity even happen nowadays? The whole "ELON MUSK BAD NOW" change to the zeitgeist is alarming to me. He was the darling of the left for years because of Tesla and SpaceX, but now he's completely persona non grata for...reasons? It smacks of excommunicati…
How you can build a successful car company in the US without getting super rich, I don't know.
Earlier quoted context omitted.
For long form, I’d suggest cold-takes blog who is very systematic thinker and has been focusing on agi risk recently. https://www.cold-takes.com
I see a lot of "we don't know how it works therefore it could destroy all of us" but that sounds really handwavy to me. I want to see some concrete examples of how it's dangerous.
Basically immediately after release both models were "jailbroken" in ways that allowed them to do undesirable things that OpenAI never intended whether that's giving recipes for how to cook Meth or going on unhinged rants and threatening to kill the humans they are chatting with. In the AI Safety circles you would call these models "unaligned", they are not aligned to human values and do things we don't want them to. HERE is THE problem, as impressive as these models may be I don't think anyone thinks they are really at human levels of intelligence or capability, maybe barely at like mouse level intelligence or something like that. Even at that LOW level of intelligence these models are unpredictably uncontrollable. So we haven't even figured out how to make these "simple" models behave in ways we care about. So now lets project forward to GPT-10 which may be at human level or higher and think about the things it may be able to do, we already know we can't control far simpler models so it goes without saying that this model will likely be even more uncontrollable and since it is much more powerful it is much more dangerous. Another problem is that we don't know how long before we get to a GPT-N that is actually dangerous so we don't know how long we have to make it safe. Most serious people in the field think making Human level AI is a very hard problem but that making a Human level AI that is Safe is another step up in difficulty.