hntrader

10 ms·

31.
by hntrader 6y ago
This is some of the reason, but I've noticed the same behavior in data scientists who are "vertically integrated" and who are only measured on the quality of the final model. In this case, their incentives are aligned with go
32.
by hntrader 6y ago
It's fundamentally a problem of incentives. If the perceived downside to giving feedback to a particular person outweighs the upside, then some selfless (or stupid) people might still decide to give that feedback, but most won't.
33.
by hntrader 6y ago
Just sort industries by political leaning. Any industry that's 40+ percent conservative might have a handful of SJWs among the engineers but they're not at the critical mass to organize and really create trouble. Wokeness gets ver
34.
by hntrader 6y ago
Politically difficult considering 30 million American potential voters own some crypto
35.
by hntrader 6y ago
Do classic lie detectors work on anyone or is it mostly pseudoscience and theatre?
36.
by hntrader 6y ago
The probability is just pow(0.5, N) where N is the student population, assuming that the distribution has no skew, assuming independence between students, and assuming the students are drawn randomly from the general population. It's a
37.
by hntrader 6y ago
They're not incorrect. Firstly, you're assuming independent probabilities between the students. If it's a single class of 30, for certain skills it might be much more likely than pow(0.5, 30) if they were all specifically co
38.
by hntrader 6y ago
We evolved to be effective deceivers, so if there was a systemic and obvious physical tell (looking left = liar) then evolution would try to eliminate it.
39.
by hntrader 6y ago
Ideally the interface is named in a way that's self-documenting and makes it clear that it's not referring to moving around your house, no downcasting required.
40.
by hntrader 6y ago
We do go to the door and call door_open(DRIVER_SIDE) - it's just second nature now because we've done it so many times that it takes up no conscious bandwidth. But that subroutine is running somewhere in the brain. To use a car, w
41.
by hntrader 6y ago
Mostly that, but I would add cultural and competence issues as well - more internal politics, and C-teamers sneaking in that drag the team down. It's easy to make sure everyone in a small team is elite and has value alignment, and it&#
42.
by hntrader 6y ago
Amdahl's Law says that the marginal benefit of additional cores is always positive (or zero). What you're describing is negative marginal benefit.
43.
by hntrader 6y ago
We need a tax on the products (or the inputs) that are responsible for this.
44.
by hntrader 6y ago
The in-game EPM ended up about 180 with spikes smaller than humans' spikes. It also had camera constraints on where it could click unlike the bot that Liquid-TLO faced. Serral has a 300+ EPM for comparison. The bot's EPM is quite
45.
by hntrader 6y ago
What was inaccurate about it for those that know little about Turing's life
46.
by hntrader 6y ago
That randomness has to be massively constrained, though. Well over 99.9 percent of inputs are guaranteed to lead to bad results. For example if we're randomly way pointing a drone, we're almost guaranteed not to be sending it some
47.
by hntrader 6y ago
It worked in adversarial board games (Go, Chess) and poker. We now have unexploitable bots for these games. It hasn't worked yet in Starcraft because the strategy space is so much larger and the action space is also much larger. The ne
48.
by hntrader 6y ago
He didn't get to use his own hotkeys? Are we sure about that? I'm sceptical, it'd make the game unplayable and it's easy to import hotkeys.
49.
by hntrader 6y ago
Isn't the general method of prevention just to train a bigger model for longer, so that all these niche edge case exploits get found and addressed during policy exploration? If there's an exploit that's sufficiently rare and
50.
by hntrader 6y ago
This is not accurate. AlphaStar was nerfed quite heavily to achieve near APM parity with humans. The early version that beat Liquid-TLO had superhuman spikes in APM (despite having the same mean APM) but they addressed that later. The bot&#
51.
by hntrader 6y ago
Self-selection bias. The insecure, the underachievers, the lonely, the narcissistic, etc, might be more likely to join. The confident, the successful, the popular, etc, probably are less likely to join.
52.
by hntrader 6y ago
They've moved beyond that now. Shadow Fiend won mostly due to superior micro but their newer 5v5 bot has human-level "planning" and executive function.
53.
by hntrader 6y ago
Well, it literally is a conspiracy theory. Whether it's a true or false conspiracy theory is what's being discussed.
54.
by hntrader 6y ago
Homes aren't a natural monopoly. Highways largely are. "there is not enough available/affordable land to build competition" There definitely is in the large majority of places. The only city that can possibly sa
55.
by hntrader 6y ago
Our highways are free because they're a public good (in the economics sense) and prone to monopoly pricing due to geographic constraints on competition. While it isn't inconceivable that housing could be public, the same underlyin
56.
by hntrader 6y ago
That's only currently possible to the extent that we can build a realistic simulator for the problem we are trying to solve, over which an RL model can learn a good state-action function. Which means that most real-world problems invol
57.
by hntrader 6y ago
How is it the "market restricting access" if the rents went up because of the eviction moratorium? If that's the reason, then what you said is true in a trivial literal sense, but that's blaming the consequence (market p
58.
by hntrader 6y ago
This was one of the few UI changes that I never got accustomed to over time. It's still just as bad, irritating and ugly as the first day I used it.
59.
by hntrader 6y ago
It takes painfully too long but there's an overall pressure in that general direction.
60.
by hntrader 6y ago
"the freeloader had to take the class again." If only that were the norm. Unfortunately freeloading is typically rewarded in university assignments. People say "that's real life" - well, no. In the workforce

More ›