bloaf

25 ms·

271.
by bloaf 4y ago
I think the real biggest "gotcha" is not understanding where the control loops are in your data. The classic example is the "air conditioned room" thought experiment. Imagine feeding an AI model a dataset with a bunch o
272.
by bloaf 4y ago
Maybe its secretly the Mill CPU architecture!
273.
by bloaf 4y ago
> Corporations have no interest in producing systems with the relevant properties to have "rights". AGI isnt of interest to business. Business is interested in automation; intelligence in this sense is precisely its opposite: a
274.
by bloaf 4y ago
I notice you've not cited it either. https://journals.sagepub.com/doi/abs/10.1177/001698629303700... > Results indicated that parents perceived that their children were experiencing a decline in energ
275.
by bloaf 4y ago
Why do you feel like not being productive all the time is a moral failing?
276.
by bloaf 4y ago
The problem is that most things don't have specific antonyms. The "not a cat" label could be applied to literally any image without a cat in it.
277.
by bloaf 4y ago
I strongly suspect the inability to do negation comes from the fact that the training dataset contains virtually no images tagged with negated descriptors. It wouldn't surprise me to see it correctly handle negations that might be mor
278.
by bloaf 5y ago
Deep inside some large enterprise company: Jr Engineer: "Hey, I know we've always managed our little dotnet application via email and shared-network-drive, but I've been reading about a thing called "git" that we sh
279.
by bloaf 5y ago
I agree, certainly not all academics live up to that virtue; but it is still a virtue of academia.
280.
by bloaf 5y ago
> And I'm disagreeing and saying that, unless you have some sort of prior agreement with the people involved (which social media/the internet/readers of your blog don't have), you don't get to avoid consequences
281.
by bloaf 5y ago
The courts, to this day, still believe in 'truth through debate.' Science does too, although the debaters are generally expected to be working towards developing empirical ways of settling their disagreements. I believe it was Ge
282.
by bloaf 5y ago
Now I want a MIT--1 license that prohibits attributing the code to the author in any way, including in legal filings, and then you don't need any other part of the license.
283.
by bloaf 5y ago
Sure, obviously you can't defy physics just by getting smarter. My parsing of the singularity is the point at which the speed of improvement exceeds our ability to understand it. "Predictability" is the usual term, but I thi
284.
by bloaf 5y ago
Employers would pay apartment owners to keep a below market apartment "available" but turn down all applicants, then use that below market price to justify arbitrarily low wages.
285.
by bloaf 5y ago
In my mind it's just a colloquialism for a certain kind of illiberalism: > "Cancel culture" is the belief that certain beliefs about race, gender, and equality are so utterly indefensible that anyone holding them is nece
286.
by bloaf 5y ago
Yes. The video fakes are always blurry around the mouth and gestures don't match the words. The audio fakes are too "clean" (too defined breaks between words and no background noise) and sound like caricatures. Text is virtua
287.
by bloaf 5y ago
Max Miller has a series about making historical foods, and references this book regularly: https://www.youtube.com/c/TastingHistory
288.
by bloaf 5y ago
I also wonder about the "Putin loses power in Russia" scenario that the author brushes aside. It seems to me that we could fight Putin on the "internal politics" front in a way that A) forces a choice between escalation
289.
by bloaf 5y ago
As a bit of a weird hobby, I like to read up on right wing conspiracy theories. That means I do a fair number of searches for specific terms and people mentioned in fake-news facebook/forum posts. Google seems to slowly oscillate betw
290.
by bloaf 5y ago
“How did you go bankrupt?” Bill asked. “Two ways,” Mike said. “Gradually and then suddenly.” Now I certainly agree that there is no realistic scenario where any significant portion of people can just choose free fish in 5 or even 10 years[1
291.
by bloaf 5y ago
If you give someone a fish, they starve after a day. Teach someone to fish and they can eat + sell the fish. But if I make a fish-catching machine to catch fish for everyone, does everyone eat, or does everyone starve? UBI is essentially a
292.
by bloaf 5y ago
Don't forget the flawed test kits the CDC sent out at first: https://www.npr.org/2020/11/06/929078678/cdc-report-official... I agree that the link between online misinformation and real world harm
293.
by bloaf 5y ago
I kind of agree, its a bit too fragmented. I think there are a handful of "obvious" interactions that should have been called out. Politics + Space: Greater international competition for space-resources will likely begin, and we m
294.
by bloaf 5y ago
My understanding is that this article understates how far and quickly these viruses spread in the wild. https://www.nytimes.com/2021/10/14/science/bat-coronaviruses...
295.
by bloaf 5y ago
The White Man's Burden is definitely a dangerous poem, it's bad to uncritically spread its message of simple cultural superiority. But lets not pretend that it is some off-the-rails racist screed. We still do try to export our cu
296.
by bloaf 5y ago
I don't think you've actually understood. As we can see from the title, what matters is not the actual stats, but what people perceive. So if the system works silently and fails loudly, people are going to think it is more broken
297.
by bloaf 5y ago
The fundamental rule of conservatism is this: >There must be an in-group that the law protects but does not bind, alongside an out-group that the law binds but does not protect. That is why conservatives are so prickly about being regula
298.
by bloaf 5y ago
Still getting errors in Houston edit: some streams back up, chat still buggy as of 09:55 local time edit2: appears to be back ~10:00 local time
299.
by bloaf 5y ago
You should always want to have the underlying code available. Without the exact procedures they used to process their data, the only kind of "using their conclusions" you can do is the superficial "take it at face value&quot
300.
by bloaf 5y ago
> the non-organic feed would be contaminated with pesticides/herbicides. https://www.agdaily.com/technology/the-list-of-pesticides-ap... Organic farming uses herbicides and pesticides too.

More ›