2,964 karma · joined September 19, 2015
I wish they'd teach with modern English translations of Shakespeare in high schools. Maybe then kids would like it a lot more. But it seems like it's taboo to read Shakespeare in anything but the original.
Even if you were certain that AI would take over virtually everything, the problem is deciding which AI companies to invest in. Thinking back to 1996, just when this new thing called "the web" was gaining tracking and assuming you were certain it was going to be huge, what companies would be the best investment? A lot of the companies doing web stuff went nowhere. Many of the biggest successes didn't even exist in 1996. It's not obvious which (if any) AI companies are worth investing in even if you're positive that AI is the future; the current companies might be massively overvalued; the best AI companies might not even exist yet.
username words karma
1. tptacek 4,310,896 416351
2. jacquesm 3,841,209 237961
3. ingve 2,273 215283
I consulted the Bureau of Labor Statistics inflation calculator:
$600 in 1925 would be $11,264 today
$600 in 1935 would be $14,329 today
A lot of money, but I've heard that it can easily cost $10-20K today to erect a couple of poles to bring power a hundred feet to your property in a rural area these days. Do you know what distance was being covered to bring power to your grandfather?
A ten-fold productivity gain in legal services sounds simply awful for society. Imagine the time and money sink if everyone can sue you for every frivolous thing because AI can prepare and file the paperwork instantly without needing a lawyer. You'll need your own AI to defend against the onslaught of legal disputes.
Every contract for jobs and every terms & conditions for services will be 10x longer because AI has a much higher complexity threshold compared to a human. My belief is that one reason tax returns became much more complicated in the last ~30 years is because of tax preparation software. In the era of paper tax returns, there was a limit to the complexity that an individual or even an accountant could handle, so there was a limit on how complicated the government could make it.
Most normal people rarely need a lawyer in their lives. With AI's productivity explosion in the legal services, you're going to need legal services every day. Your neighbor wants to borrow your chainsaw? Your AI legal agent will negotiate a liability waiver with his AI agent.
0 to 800,000,000 in 3 years?
The fastest adoption of a product or service in human history?
I can accept a facial scan, but I draw the line at fingerprints and more invasive biometrics.
Two questions jump to mind:
Why isn't this the default in Firefox?
What is the downside? I.e., what can break by enabling this parameter?
Remember the really old days when air miles were awarded solely by distance flown rather than by dollars paid? This made no business sense. It meant that someone who flew the cheapest tickets could rack up as many points as a last-minute first class business traveller who spent massively more ticket.
With the airlines I’m familiar with, it seems that pricing anomaly has been corrected. Air miles are much more correlated with the price of the ticket these days. Eg., you don’t even get air miles on the cheapest tickets on one airline I know.
But I still wonder why the airline industry created an air miles formula so disconnected from the value of the passenger in the early days.
[1] I've verified this for myself with careful record keeping over long periods of starting/stopping different BP meds, but I'm not entirely sure it's true for everyone.
Maybe it should have been titled, "How to model the right battery choice for your application" or "Understanding trade-offs in battery design".
Didn't Apple make a great deal of their revenue in the early days (1976-1985) by selling to schools? I seem to recall that schools were their main focus. A $3000 Apple II in 1980 would be $12,000 in today's money, and schools often bought many units. Is Karpathy's product really that different from an Apple Computer in how it would be marketed to schools?
> From the linked article: First, the hackers were able to record a conversation between Dubé in his office on Parliament Hill and our Radio-Canada colleague Brigitte Bureau, who was sitting at a café in Berlin.
So many questions! Does this still work? On any phone using SS7? Including landlines and mobile phones? From anywhere in the world? To anywhere in the world? What are the limitations of the attack? Why isn't this a vastly bigger problem if anyone can listen in on anyone's calls?
How was it possible that I could take 3 years of high school biology and not have heard that one lifeform absorbing another lifeform was responsible for these amazing new capabilities? We learned about mitochondria and chloroplasts, but in a very dry way. Primary education could be so much more interesting to kids with context like this.
https://www.marinetraffic.com/en/ais/home/centerx:1.7/center...
"Since December 2004, the International Maritime Organisation (IMO) requires all passenger and commercial vessels over 299 Gross Tonnage that travel internationally to carry an AIS transponder that include a GPS receiver which collects the vessel's position and movement details."
Counting across the longest segment of ships at the narrowest point of the English Channel in the feature article I see about 18 ships. In the live image from marinetraffic I see about 8. So even though the feature article is a composite image, the English Channel is indeed pretty crowded.
To flesh out some numbers: $560k minus $100k in taxes, minus another $100k for living comfortably per year, leaves $360k that you can save each year. If you earned 4% on that savings, you have about $5M in 10 years. If you moved somewhere cheap worldwide, you could live lavishly spending $100,000 a year for 50 years. I say lavishly because you are living someplace cheap and this is $100k after tax.
(Assuming no extreme inflation or hyperinflation, and that your funds remain invested.)
"There is ongoing debate about to what extent fan fiction is permitted under contemporary copyright law. Some argue that fan fiction does not fall under fair use, as it is derivative work. The 2009 ruling by United States District Court permanently prohibiting publication in the United States of a book may be seen as upholding this position." [1]
Unless we can greatly reduce the length of copyright terms and/or clarify that fan fiction is allowed, you might as well create only original works so your months or years of work doesn't go to waste.
Update: From the link another HN user found, it says, "The Bartkira organizers knew that Matt Groening had seen a copy but to date have never received a cease and desist order. As an extra defense, Harvey decided to donate all of the book's profits to charity [including] Save the Children, a charity preferred by the late The Simpsons co-creator Sam Simon."
So the creators are walking on eggshells, making sure they don't earn a cent and hoping the copyright holders don't sue one day.
Can someone who understands the original paper give an ELI5 on how that's possible?
I understand how labeling an image of a cat as "dog" could poison a dataset.
I also understand how adding images of a toaster (which sort of looks like a handbag) to a dataset of handbags could poison a dataset.
But I don't understand what's happening in Figure 6 in the original paper. The pairs of pictures of the dogs, cars, etc. look absolutely identical. What exactly is happening there?
Why not? I thought that the illustrations fit the article better than the bland and irrelevant stock photos people usually use.
As an aside: I figure that in about 3-4 years, no one will be able to assert that they do or don’t like AI generated art because there won’t be any way to know if it’s man made, machine made, or reality.
I have an anecdote on how people get falsely labeled as having a penicillin allergy: My grandad with a “penicillin allergy” was treated a few years ago with a high dose of penicillin for 4 days without any problem. The hospital didn’t have access to his medical records and also overlooked what we wrote on the admission forms, and went ahead with penicillin as the best treatment. I asked my grandad about this allergy. He said that 40 years ago he had an upset stomach after taking penicillin and the doctor told him he was allergic. He’s been repeating that information to doctors and pharmacists for 40 years.
Another person quipped that the two medical errors--false allergy and treatment despite "allergy"--cancelled each other out!
He retired in 1996.