On Friday I was converting a constrained solver from python to another language, and ran into some difficulty with subsituting an optimzer that's a few lines of easily written Scipy; but barely being supported in another language. One AI tool found this out and fully re-implemented the solver using a custom linear algebra library it wrote from scratch. But another AI tool was really struggling with getting the right syntax to be compatible with the common existing optimization libaries, and I felt like I was repeatedly putting queries (read: $) into the software equivalent of a slot machine that was constantly apologizing for not giving a testable answer while eating tens of dollars in direct costs waiting for the "jackpot" of working code.
The feedback loop of "maybe the next time it'll be right" turned into a few hundred queries resulting in finding the LLM's attempts were a ~20 node cycle of things it tried and didn't work, and now you're out a couple dollars and hours of engineering time.
With mathematical things you can always write comprehensive and complete unit tests to check the AIs work.
TDD (and exhaustive unit tests in general) are a good idea with LLMs anyway. Just either tell it not to touch test, or in Claude's case you can use Hooks to _actually_ prevent it from editing any test file.
Then shove it at the problem and it'll iterate a solution until the tests pass. It's like the Excel formula solver, but for code :D
You could, and hope that you understand the problem domain and numerical analysis enough to hit all the hard cases. And then you’d have expanded your codebase with lots of tests that are relevant to a linear algebra library and not to what you’re trying to do.
Or you could use existing linear algebra libraries which are highly optimized, highly tested, and have a well-understood api that’s easier to review.
And then get back to the legit hard stuff, like maybe worrying if your linear solver needs preconditioning and how to best to that. Or any of the many numerical problems people tend to face when doing this kind of work.
I’m not sure why you’d give the llm a pass on reinventing the wheel here when you definitely wouldn’t with any other dev.
I'm curious why you think it doesn't work, when there are plenty of people saying it does.
There are limitations at the moment, and I don't see many people disputing that, but it must be doing something right, and its abilities are improving every day. It's learning.
Sometimes I get the feeling a lot of antis painted themselves into a corner early on, and will die on this hill despite constant improvements in the technology.
I have seen similar things many times in my career. There was a time when everyone were very skeptical of high level languages, writing everything in assembler come hell or high water, for example.
At some point it is going to single shot an entire OS or refactor a multi-million line codebase. Will that be enough to convince you?
From my perspective I like to be prepared, so I'm doing what I have always done.. understand and gain experience with these new tools. I much prefer that than missing the boat.
And, it's quite fun and better than you might imagine as long as you put a bit of effort in.
I'd rather work with a human. Even with our flaws, it's still better than constantly being lied to by a tin can. If a junior kept delivering broken results as much as the "AI" does, they wouldn't be on my team that long.
I've encountered this before with absurdly high job requirements paired with a low-moderate commensurate salary. While on call with an HR rep for a listing I had loosely met, I deduced from their responses it was done so the company could claim their talent needs aren't being met domestically, and thus would file for more H-1B or H-2B work authorization permits. This is rife with its own issues of non-transparency and offshoring, where I likely encountered it in the later stages[0] and was being paid lip service to the process without the intent of hiring.
There's tonnes of worthless merchandise and supplements of a dubious nature which The Onion, the least expected of all possible buyers, now has to find a use for. My first suggestion would be melting down all of the 500% marked up gold bars[0] and make a one-time-run charity auction collectible for the Sandy Hook families. Or upcycling all the paper in Alex Jones' books [1] into paper mache, and use it to make globes, to really stick it to the globalists!
> As for the vitamins and supplements, we are halting their sale immediately. Utilitarian logic dictates that if we can extend even one CEO’s life by 10 minutes, diluting these miracle elixirs for public consumption is an unethical waste. Instead, we plan to collect the entire stock of the InfoWars warehouses into a large vat and boil the contents down into a single candy bar–sized omnivitamin that one executive (I will not name names) may eat in order to increase his power and perhaps become immortal.*
> On top of its journalistic pursuits, The Onion also owns and operates the majority of the world’s transoceanic shipping lanes, stands on the nation’s leading edge on matters of deforestation and strip mining, and proudly conducts tests on millions of animals daily.
No, that would be unethical, running the risk of healing the animals and helping them live longer lives without a high level of informed consent that only animals deserve.
They will instead receive a set of carefully designed nocebos.
The real supplements will be shipped needy areas in developing countries and then strategically withheld from anyone that desires them.
They probably don't want to be in the business of selling unregulated, scammy, and potentially dangerous goods. They might destroy the merch. Who knows.
There were a bunch of suggestions on the bluesky thread that they should donate samples of them to researchers so that they could figure out what was actually in the fucking things.
I don't know man, it's not like the dude caused the Sandy Hook massacre, just take this win and let the victims rest in peace. Let the Onion do it's things and cut ties.
They are re-labled existing products that are sold in other places, and unironically already-recognized, before being re-labled by InfoWars, as very high quality.
If you're gonna criticize InfoWars you have my 100% support in your right to do so, but try not to post out of your ass. This is HackerNews, not Reddit.
There was a past disclosure where lead was found[0] within an in-house product. Buzzfeed did a story about sending some products to a lab and you're right they're safe existing products[1] only with Infowars' own exaggerated marketing labeled on.
To be specific, the suit against the mother was against the mother's estate, since the mother was murdered by the shooter... like right away. The suit was settled by the estate.
The suit against Remington ended in a settlement, probably because Remington didn't want a chance in hell to set any legal precedent. The fact that the families got settlements is really a symptom of how unsettled the issue of gun control is in America. Like it's completely inane that it's fully legal to manufacture and sell AR-15 rifles to basically anyone, BUT that somehow marketing them to civilians is inappropriate. Remington settled because they just don't want any possibility of the status quo moving against them.
This is not even remotely true. I have done a decent amount of shooting, some dedicated training, and own multiple firearms of different types including AR style rifles. Your sort of rhetoric is at best disingenuous and not even remotely true.
If you have ever trained with any rifle you will quickly realize that while there are hunting oriented semi-automatic rifles out there, the minimized recoil, the high rate of fire, the lightweight nature, and all the ergonomic accessories make AR style rifles incredibly fast and easy to shoot. Using a red dot site you can fire two rounds to the chest and one to the head at 25 yards in under 2 seconds with a small amount practice and training. Minimally trained people can do the same with iron sites in under 3.
I am a big fan of the AR platform because of these reasons. They are not unique to the AR, but they are unique to a class of gun that is designed with these characteristics in mind. These are not the characteristics of hunting rifles.
Honesty is important, even if it works against your beliefs!
> I don’t know exactly what compensation they should get, but this does not seem like a healthy or sustainable way for our society to deal with tragedy.
I don't know if it's healthy or sustainable, but it definitely sounds healthier than ignoring the tragedy altogether.
Agreed. It doesn't seem like a long-term solution, but it is the best way we have _right now_ to visit consequences on people/orgs that enabled the tragedy. If our society sees everything in cost/benefit, then increasing the costs of actions that lead to tragedies like this is one of the best things we can do.
They did, if even indirectly. Just like how McDonald's holds some responsibility for the obesity epidemic.
The company that makes rifles makes them to be sold. It is in the company's best interest that as many mass shootings happen as possible. By providing guns, they DID contribute to the tragedy. We can tell, because if they had never produced that gun then it would've never shot anyone.
This doesn't even touch on the fact that the reason gun laws are so lax is because these companies lobby for it to be so. Again, they are incentivized to cause as many people to die as possible. Incentives matter. If mass shootings were the next blue jeans, these companies would quickly overthrow Apple.
Blame is very hard and tricky, but any institution or system in place is responsible for an intuitional failure. And that's what mass shootings are - an institutional failure.
> Not suing others for millions or billions and spreading misery. Nothing can bring those kids back.
> Maybe the government could have offered education and employment guarantees to the families?
The lawsuit wasn't about responsibility or compensation for the school shooting. It was about the years of harassment and death threats that the families of those killed had to endure from people who believed the lies that Alex Jones repeatedly told about them.
> Not suing others for millions or billions and spreading misery. Nothing can bring those kids back.
How about not slandering the parents of the victims causing Jones' followers harass and threaten them? He could have admitted he was wrong (which he only did finally at trial and under oath - far too late), but chose to double down. What about that misery?
Jones is not a victim here. He chose greed, but got owned. The motives of the families, lawyers, etc are whataboutism at best. You're essentially arguing that if somebody throws a punch at another person, said person has no right to hit back because hitting back won't take away your black eye.
> but this does not seem like a healthy or sustainable way for our society to deal with tragedy
I don't know, this, to me, is the proper set of incentives. Nobody wants to lose money, so you better do everything you can to prevent these tragedies. If we just sob a little and move on, the systems in place will not change.
I had the goldbugs and silver bugs in mind- they'd be more than willing to pay exorbitant markup, with the feel-good ennui of it going towards a good cause. These were $100 for a 1/10 gram at the time of writing and now are sold out. Coincidence???
How the family found out about satellites is very interesting to me- such an unexplained phenomena wouldn't make sense to some groups (ie, flat-earthers deny their existence altogether), where they would have observed Sputnik zooming across the sky while only being visible at dawn and dusk. With such perfect recurrence the Lykovs would have been able to deduce that the object followed ordinary orbital mechanics against the backdrop of the celestial cosmos rather than being a supernatural object, especially that Sputnik was only in the sky for a couple months in 1957. Later ones would have have likely indicated they were a product of man; I'd personally ascribe it to being supernatural in origin!
I think this is a fabrication by the journalist. Overall, it seems to me that there was an ideological agenda behind this story. In the USSR, no topic could cause a stir in major media outlets without an ideological directive.
The family would have heard about airplanes, they were from Perm Oblast, which isn't exactly remote. It isn't exactly a big leap to think that what they were seeing was the lights of airplanes. Ships carry lights for safe navigation, and if you fly at night you also carry lights for safety.
i sometimes wonder what it was to live in early human times and make sense of everything: time, phenomena etc. i am not religious but i guess back then i would have ascribed anything to gods
I'm no anthropologist but religion must've started in parallel with verbal communication, storytelling, various degrees of higher level thinking.
That said, cave paintings, which are arguably the earliest "documentation" we have of human activity and their thoughts, depict pretty tangible things; pictures of animals, hand prints, people hunting animals, plants? that kind of thing. The earliest religious symbols may have been venus statues, but it seems that it cannot be concluded definitively whether they were objects of worship / depictions of deities. That said, there's clear signs of shamanic religion dating from the upper paleolithic, 50.000 years BC and onwards. https://en.wikipedia.org/wiki/Prehistoric_religion is a good read.
Anyway, ~52.000 years ago is still fairly recent, given homo sapiens emerged ~300.000 years ago and the earliest stone tools were from ~3.3 million years ago.
Personally, I think religion / the concept of gods or a higher power has been a part of humanity for as long as we've had the ability to think smort thoughts and communicate with one another. Some would be grounded in reality - good weather means good times - but others in myth, like the stars/moon, thunderstorms, etc.
The main hardware security bugs[0] were very low hanging fruit associated with taking over the boot chain at ring 0- it's more likely that Nintendo themselves were in a rush to get the product on the market after the perceived failure of the Wii U. Even with a secure software stack, people found a way to defeat the Xbox 360 hardware[1] by physically drilling into a chip that enforced a software lock, and George Hotz became known for his work in finding ECDSA flaws in the PS3. Companies can design these locks to last for a few years of a console's lifespan, but I think people now are determined enough to dive into these difficult problems that they're unlikely to be secured forever.
There's a reason why you have to go back to the 360 and PS3 for those examples, Sony and Microsoft stepped up their hardware security dramatically after that generation. Neither the PS4, PS5, Xbox One or Xbox Series systems have ever been compromised via hardware attacks, and those earlier ones are over a decade old now.
The Xboxes have held up extremely well on the software front as well, and although the Playstation software isn't so robust (they use FreeBSD and routinely get owned by upstream CVEs) their secure boot has never been broken, which limits how much you can do with a software jailbreak. PS3 jailbreaks had continuity where you could upgrade an exploitable firmware to a non-exploitable one while retaining a backdoor, but the PS4s secure boot put an end to that.
That's not the only reason, Microsoft and Sony did improve their security a lot but their console are also much less juicy targets than in the past as well. The Xbox and the Playstation have way less exclusive games than in the past and the difference with the PC is much smaller nowadays
I mean, it is a classic example. If you have access to the hardware and the dedication to do so, you could break almost any security. That's a hilarious example to physically drill into a chip, though
This could be “famous last words”, but as someone who has worked with chip security I’d be very surprised if anyone breaks this generation of hardware at the chip level.
A decade ago the engineers designing these chips knew there were several angles of attack but there just wasn’t enough resources put into closing these holes.
Now every know angle of attack is closed. Even if you delid the chip and reverse engineer every single gate and can probe individual metal wires on the chip, it’ll still be nearly impossible to break the hardware security. Power supply and EM glitching is also protected against (can’t speak for Switch 2 but I’m speaking in general about chips going forward)
Could be bugs and mistakes that allows someone to bypass security, of course. Both in hardware and software. But I don’t think there will be general purpose angles of attack that can be used to bypass security going forward.
I think it is less that such a thing isn't possible and more that it isn't possible on "guy alone in his basement" resource and expertise constraints. And because of awful laws like DMCA 1201 if you get beyond that, or if your work becomes widely known, you will become Nintendo's new lifetime indentured servant courtesy of Uncle Sam.
I still remember how quickly this spread and people made their own versions, right about the same time as Twitch Plays Pokemon came out! I'm glad it's had such a positive impact both globally among your players and to your career. Praise helix!
Doesn't this seem a bit "low stakes" to want more disclosures about? I view it in the same light as say someone who posts makeup reviews or video game reviews on YouTube- often it isn't their primary source of income, and the item value (<$100) seems notional. Maybe it can be treated as more of a deal if a company were to attach stipulations or keywords for the reviewer to mention, at which point it should be treated as a partnership rather than a less-than-impartial review. I subscribe to independent creators on Patreon because bias is nearly impossible to avoid when there's a quid pro quo of getting a stream of free things from companies.
often it isn't their primary source of income, and the item value (<$100)
Neither of these things magically make your review more trustworthy.
Getting free stuff taints your review, and it doesn't matter if it is low value or not. And to be fair: People get their utilities turned off for these amounts. These amounts keep folks from affording medicine. Some of them simply wouldn't be able to review so many games if they didn't get them for free.
People are more likely to view the company favorably if they are getting the product for free - and are likely to review games they wouldn't otherwise review. They should reveal this sort of bias. And it doesn't matter if you are "low stakes" or not.
Many board games are funded through crowdfunding, which regularly reach 7 figures, so there is big money here. There has been controversy recently about YouTube “reviews” of active kickstarters which are actually undisclosed paid promotion, and in one case attempting to extort sponsored videos to prevent release of negative videos instead. So there’s a bit more in this than a free board game.
It was later added it to the paid copy of the game and it's similar to a 'very hard' difficulty where the normal strategies of development don't apply and you have to include DRM most of the time to break even on per-game costs. Completing it was a difficult achievement.
Tay.ai was the Gen Z large language model that was years ahead of its time. Apparently the project was continued in other languages but in a much more controlled state with a more curated corpus more similar to Watson; as opposed to being trained on unfiltered Twitter/forum input and we saw how that turned out[0].
The feedback loop of "maybe the next time it'll be right" turned into a few hundred queries resulting in finding the LLM's attempts were a ~20 node cycle of things it tried and didn't work, and now you're out a couple dollars and hours of engineering time.