zamfi

30 ms·

361.
by zamfi 4y ago
Huh, interesting. In the ML and CS theory literature I’ve only ever seen “adversarial example” used to mean an input explicitly designed to produce a specific unexpected output, not just a worst-case output that isn’t what you want. Do yo
362.
by zamfi 4y ago
This argument that it’s pretty easy to construct adversarial examples would be much more convincing if…it were not what PageRank actually is. PageRank itself initially ran iteratively and would have had exactly this same problem. In any cas
363.
by zamfi 4y ago
You are not alone. In my experience, there’s a language of mathematics that makes an almost comically strong attempt to make itself inaccessible through jargon, single-letter variable names, and an over reliance on symbolic manipulation. If
364.
by zamfi 4y ago
0. Hah, yeah. Pretty sure you’re the first person to accuse me of being anti-intellectual...I can’t say for certain, but I think this interpretation here might be on you. 1. Indeed I think you missed my point: the causality implied in your
365.
by zamfi 4y ago
...or, you could just 1) use the raw count of inbound links, weighted by the count of inbound links of the linking pages, normalize by the total number of links on that page, and run it a few iterations to get a first approximation of the p
366.
by zamfi 4y ago
Sure, I think everyone but the bullies would prefer your second approach. In the meantime, though, reputation is kind of all you have. I’ll add though that if your gossip network is only giving you “overbearing/not” signal that’s petty
367.
by zamfi 4y ago
I for one couldn't tell whether the poster you're responding to is writing sarcastically or not. ("Most people are vegan" really?) That said... > That said, you're countering a researched article, and the experie
368.
by zamfi 4y ago
There are definitely bullying, abusive people out there in academia (just as there are in industry too). I think the references thing you mention is one reason why so many people seem to put up with it. But there are also very many extremel
369.
by zamfi 4y ago
You seemed to be making the argument that without solar owning an EV may not make economic sense for the EV owner. I don't think that's related to the outcomes of the transition to an EV mandate for new sales, which is what the re
370.
by zamfi 4y ago
The state is doing a lot on all three of these fronts. Public transport is hard because construction costs are absurdly high in areas of high density, there aren’t enough people in areas with low density, and asinine restrictive zoning prev
371.
by zamfi 4y ago
OP’s assertion is unnecessary. Whether you have solar or not, for most people it won’t be hard to avoid charging between 4-9pm. Many things are possible between now and 2035, and a mandate/subsidy for landlords to install chargers is o
372.
by zamfi 4y ago
> Unless it ships from the factory with "charge only during hours X to Y" preset and turned on, Hey look, you answered your own question!
373.
by zamfi 4y ago
Grid is mostly fine. California’s electricity use varies a ton over the course of the year; in late summer the heat is most intense and peak demand is highest, roughly 50 GW. The state is facing a historic heat wave this week, and the grid
374.
by zamfi 4y ago
The thing is, the job of the digital artist of the future will include Stable Diffusion or whatever comes next. The apprentices of tomorrow will be apprentices not at creating this kind of work by hand but at using these technologies to
375.
by zamfi 4y ago
Mind blown. I don't often feel like I'm looking straight into the future, but the way Andy describes this process 100% evokes this feeling now. In some domains this will almost certainly transform the way people create art, and it
376.
by zamfi 4y ago
For anyone who doesn't want to repeat this experience: you can almost always come down in cost, by also coming down in scope -- and retain your hourly rate. I can't speak to the project above, but typically starting with a smaller
377.
by zamfi 4y ago
NYC in the ‘80s was a hard time. This is pretty much in line with my recollections…
378.
by zamfi 4y ago
That’s a different argument, my friend, a different argument.
379.
by zamfi 4y ago
Wait, Bairoch's Belgian, though. By your earlier logic, he shouldn't be trusted with this claim any more than Wikipedia or Britannica should be trusted with the claim of British origins. Seems like we'd need a non-British, no
380.
by zamfi 4y ago
To be clear, the danger is not in differing opinions, it's in overindexing on imagined motives and using those to dismiss a claim. Unpopular opinion, I guess.
381.
by zamfi 4y ago
This kind of dangerous logic is how you end up with anti-vax conspiracy theories about Bill Gates and mind control, about how one could trust Pfizer’s Phase 3 vaccine trial data they submitted for approval, etc. On the one hand, yes of cour
382.
by zamfi 4y ago
Sure, but “next time” isn’t what you said — you said, in effect, “to look sharp, focus on your tie instead of your shoes.” That has a pretty different meaning, implying that the tie matters more and the shoes don’t matter as much. It’s not
383.
by zamfi 4y ago
In general, saying "I would like to argue that being X is not about doing Y. Instead we should do Z" is going to make people, quite reasonably think that you are arguing against Y, and that people should do Z instead of Y. "W
384.
by zamfi 4y ago
My fault - I started this whole thing, and wish I hadn't. The actual point is totally reasonable, but poorly communicated. But the issue is not about the poster confusing their point, it's about them blaming the respondent for mis
385.
by zamfi 4y ago
> Instead we should focus on > And for food we should try to focus on This post contains some very odd gaslighting: your earlier post says "should"—twice—but you are claiming here that the person responding to you incorrectl
386.
by zamfi 4y ago
> If your company was structured as a co-op, yes. Of course you'd probably make them buy in at a certain level, or else their ownership would be lower than yours by enough that they wouldn't be able to vote you down (or at leas
387.
by zamfi 4y ago
This is a bit of a weird take IMO -- CEOs already need to persuade most employees (human capital) and most investors (financial capital) that their manifesto is the best option, or the employees will quit and the investors will take their m
388.
by zamfi 4y ago
I mean…maybe? But it depends on the effect size, really. And the virus thing is just one possibility among many. Even if red meat increases your lifetime chance of cancer even by a few percentage points, then one should probably spend more
389.
by zamfi 4y ago
This is only a study of raw meat consumption, which is comparatively rare among humans. If it turns out — and this study acknowledges this limitation — that the predominant causal agent here is cancer-casing viruses, that would have pretty
390.
by zamfi 4y ago
...which brings us back to 4 kkg -- 4 kilokilograms!

More ›