HN user
I'm curious, in what niches are people using Swift for new applications these days? I've enjoyed working with Swift in the past (albeit in very limited capacities), but I haven't personally come across any Swift-based initiatives in a while. I had high hopes for Swift for TensorFlow, but it was ultimately killed off.
"Will the AI be smart enough to realize the unnecessary bits, or are you just going to layer increasingly more levels of crap on top? My bet is it's mostly the latter, for quite a long time."
Dev cycles will feel no different to anyone working on a legacy product, in that case.
My experience with Reddit ads (years ago now) wasn't that subreddit-level targeting was bad—there's a reason sponsored content is such a big marketing channel, after all—but rather that the ads platform just never worked very well.
And by "never worked very well," I don't just mean "We ran ads without good results." The whole experience was just sort of confusing and underwhelming, especially when compared to other channels like FB or Google. We suspected that the majority of our clicks were bots, based on our own analytics. The targeting always felt unreliable. Support interactions were weird. In general, the platform always just felt kind of... janky.
Don't know if that's still the case now, but at least as of a year or so ago, I knew a lot of people working in digital marketing who felt the same about the platform.
In my experience, the minibar's level of use is proportional to the sobriety of the guests + their understanding of the prices.
So, basically, drunk people and children.
This is pedantry, and incorrect at that. To invest in something simply means to spend some resources and expect a material result. Energy independence, carbon neutrality, scientific progress, etc. are all material results one could hope for while "investing".
If you think another form of energy production is better, that's a perfectly reasonable objection. Twisting the discussion into a debate over the precise definition of "investment" is silly.
Can you expand on why pornography is an unsuitable focus for a scholar?
Hacker News is a largely technical community. If there was any community that would respond to your product being open source, interoperable, or written in a niche language, this would be it.
I can understand why you're puzzled, as my comment was not about mask mandates.
But in the spirit of helping, roughly 6 seconds of Googling in regards to your question turned up this https://www.healthaffairs.org/doi/10.1377/hlthaff.2021.01072
"Reasonable" is doing a lot of work there.
There were and are plenty of scientists critical of the CDC. Immediately after officials lied about masks being ineffective, prominent scientists voiced harsh criticisms. They were not harangued for being anti-science or conspiracy theorists.
On the other hand, people who fundamentally did not grasp what mRNA is, or who believed that COVID caused no more deaths than a flu, or who touted "medicines" that had no demonstrated efficacy—they were deservedly criticized. Unfortunately, the criticism wasn't enough to prevent many of them from making quite a bit of money peddling their beliefs.
No, using engines to find lines is a common way to practice (at least, that's my understanding. Obviously, I'm not personally a world class chess player.)
I was legitimately curious about this--my memory of Windows 95 is not this nice--so I looked at Wikipedia's list of software released in 1995:
https://en.m.wikipedia.org/wiki/Category:1995_software
What software specifically do you recall being "an order of magnitude more complex" than today's popular web apps?
I appreciate your experience here. If you're interested, I'd highly recommend looking into Portugal's results in decriminalizing drugs (coinciding with an enormous reduction in opioid overdoses): https://www.apa.org/monitor/2018/10/portugal-opioid
I'd also recommending looking into the UK's previous method of treating opioid addiction, commonly referred to as "The British system." Vice is hardly an unbiased source, but they serve as a good entry point on this topic imo: https://www.vice.com/en/article/yw4nnk/how-the-us-stopped-a-...
It's important to note that the systems people hold up as evidence of decriminalization's success are rarely "solely" down to decriminalization. Typically, they involve a broader "substance-abuse-as-public-health-crisis" approach. However, decriminalization is essential for such an approach to work.
This is a false equivalence.
There is censorship in both the West and China. There is also water in both a desert and a rainforest—but it would be ridiculous to say that rainforests and deserts are therefore equivalent in terms of water.
We are bitter and old, but that doesn't make us wrong ;)
I am very critical/skeptical of more or less all things crypto, but I heard an interesting point from an artist recently who had begun selling NFTs of their work. It is in line with the thinking around the Rolex example in the article, but in my opinion, more salient:
I asked them about what appeared to me as the inherent silliness of selling an "exclusive" tokenized image, which could be easily copied with a right-click for free. Their response was that this is essentially no different from print making. When you buy a print, you are buying something you have the means to produce yourself. You aren't paying for the quality of the print making materials--you could order identical prints of the image from a print shop, or if you're less concerned with quality, simply print at home.
But when you buy a print from the artist, you're paying for that signature that says "#3 of 100." There's nothing stopping the artist from printing more, and there's nothing stopping random people from duplicating the print, but we're comfortable with the idea that a signature confers a "uniqueness" to the print that makes it valuable.
I don't know if that is an argument against print making or a case for NFTs, but I found the point interesting.
Context: I own no NFTs nor do I plan to.
In the course of a normal day, an average person might interact with a dozen different ML-powered apps just using their iPhone.
- Uber/Google Maps/Waze: ETA prediction
- Gmail: Smart Compose & spam filtering
- Instagram/Snapchat/Any camera app: Computer vision
- Siri/Google Assistant: Speech-to-text
- FaceID: Facial recognition
- Facebook/Netflix/All content aggregators: Recommendation engines
- Any banking app: Fraud detection
ML's use is extremely widespread at this point. The above list is just a tiny snapshot. "AI" is term thrown around by marketers and hypemen all the time, no arguments there, but ML's usage is anything but niche these days.
That's so odd. I worked on a project involving some Chevron employees once, and we had a strange number of conversations about employer liability should someone injure themselves in the office. It's been years, but your comment reminded me. I thought it was just a quirk of that team, like one of them was just puzzlingly obsessed with it. Maybe the oil industry is just inexplicably plagued by slip-and-fall lawsuits?
I'm pretty critical of click-bait headlines in what are supposed to be more academic journals, but I don't get the frustration with this one. 'Families' is written in quotations, which tells me right from the jump that the author is probably not speaking about literal family units, and the use of "family" or "family tree" nomenclature in discussing matters of genesis and inheritance is pretty commonplace.
Eh, if you boil all research in AI/ML down to the binary of "AGI or bust," then sure, everything is a failure.
But, if you look at your smartphone, virtually every popular application the average person uses--Gmail, Uber, Instagram, TikTok, Siri/Google Assistant, Netflix, your camera, and more--all owe huge pieces of their functionality to ML that's only become feasible in the last decade because of the research you're referencing.
I'd also be surprised if Jones kept a strictly disciplined financial operation running at InfoWars over all these years, such that there is no opportunity for the courts to pierce corporate veil.
The practical ability to move somewhere--e.g. to find a home, place an offer, have it accepted, and then peacefully coexist in the neighborhood.
These are all largely up to the discretion of individuals in the community.
A "right to speak with a human" makes sense to me on an industry-by-industry basis, in the same way that construction companies and restaurants have different regulatory agencies and checks.
Applying it to all businesses sounds like a bad idea, however. Financial institutions? Certainly. Healthcare companies? Makes all kinds of sense. But I don't see an ethical imperative for, say, Giphy (pre-acquisition) to provide that kind of support.
I think the point is that yes, that seems like a good solution for verifying content that purports to be released by a certain creator, but it doesn't solve the problem of deep fakes for captured footage i.e. you can prove it isn't a video that you created, but you can't prove it isn't a video someone else took of you.
The majority of large, popular datasets in deep learning are curated and hosted by academics:
No one who pays attention would have predicted hundreds of thousands of Americans dying from influenza, because in the last decade, it has never happened.
https://www.cdc.gov/flu/about/burden/index.html
The high-end of yearly flu deaths is in the low 50,000s, while the low-end is just over 10,000. COVID has killed hundreds of thousands of Americans each year.
You're not just misinformed about vaccines, you're outright lying about numbers that can be easily Googled.
I don't disagree, but I will also take debates about licensing over 30 separate "Looks cool! I've built something similar here: PLUGS PROJECT" comments.
For a long time now, Google has weighted behavioral signals similar to what you describe. "Bounce Rate" is the percentage of users who quickly leave your site after clicking. "Dwell Time" is the amount of time a user spends on a page.
There's even a cottage industry around gaming these signals. See SerpClix and the like.
I think in this situation, comparing housing to stocks isn't quite apples-to-apples. One of the appeals of owning a house is the money you no longer spend on rent.
...Did you read the article you linked to? The bulk of the article is dedicated to discussing why that number is misleading, and how it reflects our lack of data/transparency when it comes to policing, particularly when it comes to police involvement in homicide.
I promise, if FB/other companies could automate this away, they 100% would.
In general, it's easier for a computer vision system to recognize and filter a video that has already been banned--though, there is a constant arms race here as well--than it is for it to judge the content of a completely new video. That means that, for a huge number of cases, a human being will have to see the footage at least once.
EDIT: To be clear, I'm not taking up for FB in this situation. I'm specifically clarifying the difficulty of using ML/DL in moderation systems.