1:02:00 I like this argument: some people love having more power than others...they really dedicate themselves to it...Ensuring openness and democracy allow the rest of us to invest in defense...cos its much easier to work on defense than attack...and stop the attack path when such people go crazy...
As of the last few weeks ChatGPT with GPT4 has been garbage. It just like when you google a question and get a vague, off topic list of possibilities. It used to be specific.
I listened with interest, and Jeremy is an actual expert. But I am not persuaded that superintelligent AI is far away, or that an AI on every desktop offers much protection. If anything, AI for all is going to accelerate the discovery of architectures for efficient learning and agentic AI, and that's just getting closer and closer to the conditions for successfully self-enhancing AI. And superintelligent AI seems likely to overwhelm humanity even if we are assisted by AIs of human intelligence. So we still need something like "superalignment". Futhermore, even if superintelligent AI that surpasses humans-with-AI is somehow chimerical, won't humans-with-AI overwhelm humans-without-AI? And I also note that democratic processes seem to have a very small part in determining the outcome of humanity's AI adventure. AI is driven forward by money and power now, decentralization depends on the boldness and ingenuity of maverick techies, and either way, the whole thing is an enormous experiment in creating nonbiological thinking beings, potentially giving away humanity's chief evolutionary advantage to entities that aren't even DNA-based.
@@VoltLover00 Space tourism, blue LEDs, and Tom Cruise flying a jet fighter for real. They all belonged in a movie until they didn't ;-) I don't think it's particularly speculative that humans-with-AI could overwhelm humans-without-AI, since that is already the case with many forms of technology. For that matter humans-with-money in general overwhelm humans-without-money, and AI is just one of the techniques they might use to achieve that. But there's a whole bunch of dangers we have to get past before we even reach the danger of AI replacing humans. We already have systems such as electrical grids that people rely on for life and that aren't always well understood (hence the 2003 blackout). Adding AI to safety-critical systems such as the electrical grid or international finance, is almost certainly going to result in at least one major foot-shooting by humanity. But to varying degrees that's how all progress works.
This was fantastic. A lot to learn from Jeremy
Q*anon I'm dying 😂
In the past two months - Mark Zuckerburg said the next Llama will be state of the art. GPU Sales also show Meta in the lead, tied with MS.
Is that a Tim tam Hugo was eating in the beginning 😂?
I think he ate 2 or 3 double coated tim tams actually and Jeremy got jealous!
1:02:00 I like this argument: some people love having more power than others...they really dedicate themselves to it...Ensuring openness and democracy allow the rest of us to invest in defense...cos its much easier to work on defense than attack...and stop the attack path when such people go crazy...
Share Spotify link, it's trouble to hear while traveling
37:37 mistral 7B
As of the last few weeks ChatGPT with GPT4 has been garbage. It just like when you google a question and get a vague, off topic list of possibilities. It used to be specific.
Could i run mistral 7B on raspberry pie ?
I listened with interest, and Jeremy is an actual expert. But I am not persuaded that superintelligent AI is far away, or that an AI on every desktop offers much protection. If anything, AI for all is going to accelerate the discovery of architectures for efficient learning and agentic AI, and that's just getting closer and closer to the conditions for successfully self-enhancing AI. And superintelligent AI seems likely to overwhelm humanity even if we are assisted by AIs of human intelligence. So we still need something like "superalignment".
Futhermore, even if superintelligent AI that surpasses humans-with-AI is somehow chimerical, won't humans-with-AI overwhelm humans-without-AI? And I also note that democratic processes seem to have a very small part in determining the outcome of humanity's AI adventure. AI is driven forward by money and power now, decentralization depends on the boldness and ingenuity of maverick techies, and either way, the whole thing is an enormous experiment in creating nonbiological thinking beings, potentially giving away humanity's chief evolutionary advantage to entities that aren't even DNA-based.
Any scenario that sounds like it belongs in a movie is completely unrealistic
@@VoltLover00exactly. It's clear some people never programmed before because building anything sufficiently large of sophisticated is extremely janky.
@@VoltLover00 Space tourism, blue LEDs, and Tom Cruise flying a jet fighter for real. They all belonged in a movie until they didn't ;-) I don't think it's particularly speculative that humans-with-AI could overwhelm humans-without-AI, since that is already the case with many forms of technology. For that matter humans-with-money in general overwhelm humans-without-money, and AI is just one of the techniques they might use to achieve that. But there's a whole bunch of dangers we have to get past before we even reach the danger of AI replacing humans. We already have systems such as electrical grids that people rely on for life and that aren't always well understood (hence the 2003 blackout). Adding AI to safety-critical systems such as the electrical grid or international finance, is almost certainly going to result in at least one major foot-shooting by humanity. But to varying degrees that's how all progress works.
was there a name for that technique jeremy mentioned for the good model training the "shit" model?
Distillation?
Great talk mostly though the bellyaching was a bit cringe….
I mean he is trying to make LLMs uncool so being mid and cringe is right on target.
Hi, really enjoyed the talk.
Does Jeremy has an X account I could follow?
Thanks
@jeremyphoward
the guy just hates open ai
you love it ?
that’s one way to categorize what you don’t like hearing
Me too, mostly moron ping AI search engine