staunton

19 ms·

301.
by staunton 2y ago
> this is turning into huge power/money grab The power grab happened a while ago (the shenanigans concerning the board) and is now complete. Care for humanity was just marketing or a cute thought at best. Maybe humanity will survive
302.
by staunton 2y ago
Maybe it should not be done but the large neutral networks this decade absolutely rely on this. A network at the global minimum of any of the (regularized) loss functions that are used these days would be waaay overfitted.
303.
by staunton 2y ago
In practice, you don't want the global optimum because you can't put all possible inputs in the training data and need your system to "generalize" instead. Global optimum would mean overfitting.
304.
by staunton 2y ago
I broadly agree. There are caveats, however. Things might develop a dynamics of its own, not necessarily intended or in the interest of any major player. Similarly to how Facebook don't necessarily want to promote political extremism,
305.
by staunton 2y ago
It doesn't matter if you call what it does "intelligence" or not. We need to understand what it can do. And I'm certain that it can be used to spread misinformation on social networks. It can even be used to spread pers
306.
by staunton 2y ago
> then 4C is unavoidable Is this estimate of yours based on anything or just gut feeling? I like to draw a comparison with world wars. As a fraction of GDP, the world spent more on WW2 than what is needed to solve climate change. Having
307.
by staunton 2y ago
Do you have a link where the BBC repeats this? I've never heard of any of it.
308.
by staunton 2y ago
We'll start trying when the consequences are felt. It will be too late to prevent a major catastrophe but not too late to prevent an even bigger one.
309.
by staunton 2y ago
> the current political regime will be overthrown You seem oddly confident in that prediction. There's a lot of alternative paths though.
310.
by staunton 2y ago
> It's too difficult for customers to organize a protest in a way that will influence that business's decision-making. It's actually very easy (in theory). You (vote for someone who will) ban fraudulent and anti-competit
311.
by staunton 2y ago
What is "Zaka" and what has BBC been saying? Have a link to what you mean?
312.
by staunton 2y ago
Lean is similar to Coq. You could - state and prove things about (e.g.) C code, like the "Software Foundations" book does for Coq. However, nobody seems to be doing this with Lean and the tools are missing. - write programs in Lea
313.
by staunton 2y ago
> My guess is it's because the problem is so much simpler. No variables, no loops, no recursion, etc. My guess: it's because people actually care about correctness for hardware. It's really expensive to throw out all those
314.
by staunton 2y ago
> those machines that used to be brutally efficient at doing what you asked of them. In a way, that's a natural thing to happen. Computer systems are weird in that we have an extremely good model of how they work at a certain level:
315.
by staunton 2y ago
> It's always going to be prone to corruption, fraud, and mismanagement. A lot of things get done despite corruption, fraud and mismanagement. As soon as people decide to actually do something about it, it will be done. The amount
316.
by staunton 2y ago
Like who?
317.
by staunton 2y ago
> most applications have broken implementations What applications? Almost nobody writes their own implementations of encryption nowadays (nor should they). You mean openssl is "broken"?
318.
by staunton 2y ago
I know lots of people who are authors on papers they've never even looked at. (They implicitly agreed to be authors and know that there's some such papers, they just didn't even read the tiles)
319.
by staunton 2y ago
Using neutral networks to pre-filter data, especially networks trained on stimulations, should be seen as somewhat suspect as it potentially introduces bias into the analysis that is very hard to quantity, let alone control. Unfortunately,
320.
by staunton 3y ago
We have "enough time" to explore stuff. You or me will not live to see it but what's stopping a spaceship from going to some place for a few million years?
321.
by staunton 3y ago
> So what you are suggesting is that we do away with that (fine!) and replace it with --- something handwavy (not fine). I wasn't really trying to suggest any concrete system to replace the current one. Neither would I be able to do
322.
by staunton 3y ago
> my proposal of paying reviewers and replicators addresses [all concerns] Your other comment didn't say who might be paying reviewers. Journals clearly won't (why should they, they have grrat profits in the current system an
323.
by staunton 3y ago
Like I said, it's not clear to me what exactly established journals have been doing "better" historically than a preprint server could. You say they are better at "being a journal" -- OK. They are established, wel
324.
by staunton 3y ago
I agree with you but still want to play devil's advocate regarding you argument: > [It's easier to write specifications than implementations because] in a specification I can specify outcomes without having to say how they&#x27
325.
by staunton 3y ago
Why do you name Lean here as opposed to the other provers? Lean is "the new shiny thing" but my impression is that the community sees its focus in math, not so much software verification. Quite a few design decisions (e.g. focus o
326.
by staunton 3y ago
What are those reasons? The only thing I see that journals do which preprint servers couldn't easily take over is prestige .
327.
by staunton 3y ago
> Even if we assume memories/repeaters to exist (iiuc this should contradict the no-cloning-theorem) you’d need to trust them not to listen in The whole point of having repeaters is that it's impossible for them to listen in.
328.
by staunton 3y ago
> The world is so interconnected and everyone depends on everyone so much that there is basically no point in planning for this scenario This sentiment was very prevalent right before WWI.
329.
by staunton 3y ago
> manipulating other people. [...] It's why psychology was founded in the first place. What are you referring to here? Which specific founder(s) wanted to (or did) manipulate people?
330.
by staunton 3y ago
You said > the GHZ experiment [1] can rule out local hidden variable models and confirm QM predictions with no statistics at all However, one needs to use statistics to even show GHZ works. That does sound contradictory to me. The correl

More ›