digbybk

9 ms·

31.
by digbybk 2y ago
> There are hours of podcasts with Chomsky talking about LLMs I'm not an expert, but it seems like Chomsky's views have pretty much been falsified at this point. He's been saying for a long time that neural networks are a
32.
by digbybk 2y ago
I'm not more knowledgeable, but can't you square it with the fact that a for-profit arm is a new direction, and when organizations make big structural changes like this there are often people inside the org who disagree and decide
33.
by digbybk 2y ago
Particle accelerators are pinging the deepest layers of reality that we can possibly reach. Deeper than anyone could have imagined just a few generations ago. That anyone can be cynical about that is hard to understand.
34.
by digbybk 2y ago
I think the Future is a massively under appreciated album.
35.
by digbybk 2y ago
> Rollercoasters and psychedelics can be a great pairing Reading this sentence made my arms tingle with anxiety.
36.
by digbybk 2y ago
Coffee being the only exception.
37.
by digbybk 2y ago
I was disappointed when I realized that Nextdoor was an even worse place than other social media platforms. In principle it should be like FPF. Is it the moderation or does FPF work because Vermont is so sparsely populated and homogeneous?
38.
by digbybk 2y ago
Is that dissonance or shows that the concept is generally applicable? Human inventions can be misaligned with human values. The more powerful the invention, the more damage it can do if it is misaligned. The corporation is a powerful invent
39.
by digbybk 2y ago
Can you elaborate? How are the ethics different from any role in tech?
40.
by digbybk 2y ago
> This is a problem of underutilisation in my view. Too many properties are being used as investments and not as a primary residence. I hear this argument a lot and always wonder if the causal direction is flipped: people use housing as
41.
by digbybk 2y ago
I mostly notice it only when I meditate. I also notice that my nose hurts.
42.
by digbybk 3y ago
Repair and widening are different things. When you say we're nowhere near saturating our landmass with people, are you thinking of the landmass of the entire country or the areas within an hour commute or less of a major population cen
43.
by digbybk 3y ago
> I like to think my opinions are not that bad. I find this sentence kind of funny. If you thought your opinions were bad wouldn’t you change them?
44.
by digbybk 3y ago
What’s the relative likelihood that a waymo car will hit a cyclist or pedestrian vs a human operated car? Unfortunately for us cyclists, people are hit by human drivers every day.
45.
by digbybk 3y ago
You’re certainly not the only two. https://en.wikipedia.org/wiki/Dark_forest_hypothesis
46.
by digbybk 3y ago
I really enjoyed Seveneves.
47.
by digbybk 3y ago
I’m not claiming otherwise here. I’m claiming that it’s not a good test because a human would (almost certainly correctly) also skip past the double feather and answer the likely spirit of the question.
48.
by digbybk 3y ago
I wouldn’t even say it’s failing. Recognizing and correcting obvious errors in speech is a feature not a bug, both in humans and LLMs. Is it more likely that they meant to say feather twice or they mistyped?
49.
by digbybk 3y ago
We have an evolutionary aversion to situations and things that may lead to our deaths. Staying alive is a “terminal goal” for most humans, not an instrumental goal (with the exception of people who seek to commit suicide as you point out, b
50.
by digbybk 3y ago
https://youtu.be/ZeecOKBus3Q Basically, any sufficiently advanced agent will have “prevent anyone from switching me off” as an instrumental goal. And any sufficiently advanced intelligence will come up with ways of achievin
51.
by digbybk 3y ago
> It should be obvious: power. But that also means that it doesn't matter what they say, because they'll lie to conceal that this was really a power play, driven by petty looking rivalries. Am I the only one who finds these kin
52.
by digbybk 3y ago
> The only people who truly believe this AI safety stuff are those with no technical knowledge or expertise. This is a demonstrably false claim that gets repeated over and over. 2 out of 3 Turing Award recipients for work in AI are peopl
53.
by digbybk 3y ago
Money and power is still not a satisfying explanation. If everything had gone according to plan, how would be have ended up with more money and power?
54.
by digbybk 3y ago
https://aisafety.info/
55.
by digbybk 3y ago
So either doomers are wrong about the technology, or we’re doomed. There’s no universe where doom is possible but for policy decisions and we aren’t doomed.
56.
by digbybk 3y ago
Our future selves will always have more information than our present selves, so we should let them make the decision whenever possible.
57.
by digbybk 3y ago
This reminds me or the joke where an economist refuses to look down when told that there’s a $20 bill on the ground: “if there was really $20 on the ground, someone would have picked it up already”. Seriously though, we’re just getting star
58.
by digbybk 3y ago
Iirc the robots are the equivalent to land mines, continuing to cause harm after the war is over. That episode felt very plausible and horrifying.
59.
by digbybk 3y ago
There’s a distinction between pleasure and happiness. - why cultivate friendships, have experiences like travel, build hobbies and skills that will give you joy all through your life? For many people, the time and financial cost of children
60.
by digbybk 3y ago
It is trained to predict tokens but to do that well it needs to develop some internal structure. That structure is mostly opaque but it appears to model the world increasingly well with scale.

More ›