jxcole
jxcole on Hacker News
It's probably caused more by anthropogenic climate change than sun luminosity.
I think the point here is that LeCun was arguing that training on pure text would not grant spatial understanding. I believe most models are trained on spatial data as well, so you are both right.
If many people hated their products, they wouldn't use them. If they didn't use them they would not be financially successful. Google is not dying just because you personally have a feeling the results are worse than before and there is no definition of dying that would be consistent with Google's current state. If Google were to die they would die the way Yahoo did, because a competitor was demonstrably better than them and everyone switched off. There is no realistic evidence that this might occur in the near future.
The obvious solution is to just copy paste it into Claude itself and ask it to fix. Works for almost any Claude problem
A better analogy would be to compare it to being able to install anything from online vs only installing from an app store. If you wouldn't trust an exe from bad adhacker.com you probably shouldn't trust a skill from there either.
IANAL but I would be scared of getting sued. For example, if I try to give a perfectly good car seat to good will they refuse to take it for liability reasons. Baby safety is serious business.
He would need to fire at least 30% of the staff to make a difference. You can't fix the rot that started with the McDonald Douglas purchase by changing one person.
If you look carefully you see that they actually started with a much bigger drone and then shrunk it down. The minimum is 9 mg (that they could build) but it sounds like the design can be expanded to almost any size. A more obvious question would be is there any functional difference between this and a balloon.
Each pilot can have a button they must always press, if both are released loud tritones alarms and flashing will fill the cockpit, something no one can sleep to.
Surely any one with 700 million users could just build their own?
I disagree completely. Facebook was just Myspace without the terrible UX. Google was just Yahoo but better. Tech businesses can easily move in over each other if there is a way to make it noticeably better.
I think it's possible that some generous person determines a sequence of prompts that generate, say novels, and then pipes these prompts into a program causing thousands or even millions of wholly varied novels to be generated in the public domain. I imagine this is what the OP meant.
Why would that make it less reversible? Honestly it's not like it was ever easy to reverse DNA damage. Reversing other kinds of damage might be harder, but unless you have more information it's not clear that it _must_ be.
This seems to be Hinton's MO though. A few years back he ripped out convolutions for capsules and while he claims it's better and some people might claim it "has potential", no one really uses it for much because, as with this, the actual numerical performance is worse on the tests people care about (e.g. imagenet accuracy).
https://en.wikipedia.org/wiki/Capsule_neural_network
"everyone depends on their particular bolt, and we are all screwed."
I love clever puns.
For the curious: it's not intentionally NSFW but the filters aren't very good (if there are any) so you do see occasionally inappropriate images.
Apple's stance on air tags is ridiculous. For one thing, you will only know if someone is criminally tracking you if you are an apple user. Android users have no luck.
Secondly if this is true and you stole a bag, all you need to find the air tag would be to hold an iphone close to it for a few minutes. So what's the point of it then?
This article misses the point completely. Fundamentally the question is, if some technology has made us twice as productive, why don't we work half as much? This will never happen because capitalism and economics are completely psychological. You could look at any period of history and see some invention that caused a 2x increase in productivity. Inventions like farminng, the electric or gas powered motor, etc. Humans will always fill the void left by their extra productivity with more work.
We have the technology _today_ for the average work week to be about 5 hours we just choose not to do this as a society.
I bet it was a co worker who wanted to leak but also didn't want to be blamed. Or maybe even a coworker who had a grudge against the defendant and didn't care about the leaks at all.
It would be pretty easy to set up. If you work in the same room or building as a coworker how hard is it to set up a camera or a physical key logger to steal their password? Once you have someone's username and password you can make it look like they did anything. You could even do something nefarious on their computer when they went home for the evening.
You may think that as security professionals they would definitely notice a key logger, but do you honestly think _anyone_ checks the back of their computer every time they come back from a lunch break?
Having been on a jury in the US one time (attempted murder), the purpose of this system is probably not at all obvious. The idea is to put 12 people in a room and force them to agree to the same thing. You can deliberate almost any amount of time you want. If you try to tell the judge after a single day of delibrations that you are a hung jury, the judge will force you to stay longer. Only in extreme cases where the jury has been hung for a very long time does the judge allow a mistrial.
So the idea is to force 12 people to convince each other of one idea or the other.
If you're going to cheat why not all the way:
document.querySelector(".time").innerHTML = '<b>Time:</b> 0.0001';
Array.from(document.querySelectorAll(".theSport input[type='checkbox']"))
.forEach(e => { e.disabled = false; e.checked = true; });
We used almost this exact scheme for app id indices and the curious problem we had to design against was inadvertent profanity. At some point we decided to just never use vowels to avoid ever having a complaint about 12f*ck if in the URL
I feel like you are very defensive here and I want to be sure we take time to recognize this as a real accomplishment.
Seriously though, I do doubt I can be fully replaced by a robot any time soon, it may be the case that soon enough I can make high-level written descriptions of programs and hand them off to an AI to do most of the work. This wouldn't completely replace me, but it could make developers 50x productive. The question is how elastic is the market...can the market grow in step with our increase in productivitiy?
Also, please remember that as with anything, within 5 years we should see vast improvements to this AI. I think it will be an important thing to watch.
IIUC most of the mass required to build plants comes out of the air (CO2). Of course, any astronauts wanting to utilize these fungi will have to take some amount of other resources with them.
The most important measure for any radiation shielding in space is weight. We have great technology for blocking all radiation today: lead plating. This is obviously way to heavy for space. Alternative suggestions have been to surround the astronauts with water which presents a lot of challenge in terms of plumbing and weight. Having the walls covered in fungi may actually be a much lighter option, I certainly hope the idea is well considered.
> If someone is competent enough to sustain 2 jobs, they will also get quickly promoted to 2x their salary if they are working one job.
I've found the opposite is true. If you are extremely competent at your job your manager is motivated to keep you in place because that's what you're good at. It's easier to promote lower-accomplishing people-pleasers because it won't affect _their_ software throughput as much (higher level engineers write less code).
The actual paper is
https://openreview.net/pdf?id=RQUl8gZnN7O and the title is "Learning to see by looking at noise". I think a lot of the negative reactions here are from the rather ridiculous and inaccurate re-interpretation of the paper in the article.
The paper itself seems ok. It's certainly not ultra groundbreaking, but I think the research is useful and presents a pretraining step that could be used in many applications.
I certainly self censor like this on a different subject: I'm one of the only people who thinks the tech interview process is actually pretty good (I'm also very good at it, so that could be why). But anyway, only an idiot wouldn't know that a positive statement about tech interviews would garner a thousand downvotes. People seem to think HN is some sort of bastion of free speech but really it's just an echo chamber of slightly different opinions than, well, other echo chambers.