im3w1l

21 ms·

301.
by im3w1l 1y ago
It certainly seems plausible, but I wouldn't entirely rule out other possibilities. Do to give an example if you present the LLM with two people that are exactly the same except they have different color shirts I think it will suggest
302.
by im3w1l 1y ago
I think what will happen is that ordinary people will invest in a mutual fund managed by AI and/or the funds people already invest in will start adopting AI-tooling. I think this scenario is plausible because the path to this scenario
303.
by im3w1l 1y ago
Webgpu api taking screenshot of full desktop maybe?
304.
by im3w1l 1y ago
When you say cracking open a black hole do you mean cracking the event horizon to form a naked singularity?
305.
by im3w1l 1y ago
Platform 9¾ at King's Cross Station
306.
by im3w1l 1y ago
I think you know exactly what to make of it.
307.
by im3w1l 1y ago
Just minddumping here but I think that basically high velocity is bad because it leads to high friction and high temperatures. Velocity has two components, downwards and tangential. Tangential velocity is both good and bad. It increases vel
308.
by im3w1l 1y ago
Timeframes are hard to predict. I just notice these signs: That it can suggest and reason about strategies to improve itself, and break those strategies down into smaller goals. That it can integrate with toals to (to some extent) work on t
309.
by im3w1l 1y ago
I think current LLMs are smart enough to trigger the intelligence explosion. And that we are in the early stages of that.
310.
by im3w1l 1y ago
Unless you race the beam the shape will be there in graphics memory which you could read out.
311.
by im3w1l 1y ago
Before concluding a genetic change took place I would want to rule out alternative hypotheses. Maybe the cyanobacteria needed to do a mode switch which took time, maybe some parasite had to die, or some symbiote had to enter the culture. Ye
312.
by im3w1l 1y ago
This particular study looks like it's dealing with a pretty narrow condition and solution (protein missing - add gene for protein). I don't think this particular research can be extended the way you hope.
313.
by im3w1l 1y ago
To me this is backwards. I don't think there is a common need to handle a generic file not found error. Let's say the user tries to open an image file. The image file exists, but when decoding the image file you need to open some
314.
by im3w1l 1y ago
I understand what you are saying, but this seems like a very low bar for prior history. Many people are into stuff like that without ever going crazy, which contradicts your assertion they are all disasters waiting to happen. It's also
315.
by im3w1l 1y ago
A key question I think is whether these models can be made safer for mentally ill without cutting into utility for the healthy. Is the danger inherent to this technology, or is it incidental? And here I think we are fortunate that there doe
316.
by im3w1l 1y ago
Since this discussion is still active, I think hwillis was the only one that got my idea. Pumping heat into the radiators will make them hotter then they would be by just passive conduction, and then the T^4 radiation scaling means that the
317.
by im3w1l 1y ago
This sounds like it would be mostly great but occasionally really awful if you get a lot of matches.
318.
by im3w1l 1y ago
Okay this may sound silly but what about a solar powered ac for cooling? Like solar radiation is 6000K right, so if you used that to pump your waste heat into say a 1000K radiator (aimed away from the sun obviously) I'm thinking it mig
319.
by im3w1l 1y ago
My guess is that online dating in general and tinder in particular are key drivers of this. Predicting whether the trend will last is thus roughly synonymous with predicting the survival and evolution of these services.
320.
by im3w1l 1y ago
I think it may be because humans have a higher intelligence than other animals. This makes it possible for men to display genetic fitness in other ways than looks. For women, in constrast, the most important thing to signal is their capabil
321.
by im3w1l 1y ago
Adversarial compatibility is pretty nice though, I don't think we should do away with it.
322.
by im3w1l 1y ago
> google/samsung/apple/aws/cloudflare/microsoft One thing to keep in mind is those people are already paid quite well. What life can you offer them that they don't already have? Blackmail is a likelier angle
323.
by im3w1l 1y ago
They absolutely do talk about it. Maybe you should ask yourself why you never heard about it though.
324.
by im3w1l 1y ago
I think floating point is viable - balatro doesn't exactly perform billions of floating point operations per second.
325.
by im3w1l 1y ago
I think it's simpler than that. I think they hire people interested in the subject of AI safety and give them relatively free hands to publish what they find, and findings don't necessarily have to be part of some agenda that bene
326.
by im3w1l 1y ago
This but unironically. In the movie no one was killed for the purpose of being turned into soylent green, they just used bodies of people who were dead anyway. I remember thinking that it's not such a big deal compared to all the other
327.
by im3w1l 1y ago
Sure but you can ask that at different scales, in a reductio ad absurdum: Why should EU use American tech company? But also why should Germany use French tech company? Why should one region of France use tech company from other region? Why
328.
by im3w1l 1y ago
Would addition be slower than for normal floats?
329.
by im3w1l 1y ago
So my pet theory for this space is that a pure exponent format could be interesting. Like a number would be s * 2 ^ (e / C), where s is a sign bit, and e is a signed int and C is some constant maybe 2^20 ish for a 32 bit format. Specia
330.
by im3w1l 1y ago
I think most everyone can agree with this: Being 100% rigorous and rational, reasoning from first principles and completely discarding received wisdom is a great trait in a philosopher but a terrible trait in a policymaker. Because for the

More ›