tinco

It would come pretty close, probably close enough that you wouldn't be able to tell the difference on 90% of projects.

What is more expensive, copying the message, or memory fencing it, or do you always need both in concurrent actors? Are you saying the message passing overhead is less than the cost of fragmented memory? I wouldn't have expected that.

Why couldn't a machine that identifies relations between tokens be AGI? You're imposing an arbitrary constraint. It is either generally intelligent or its not, whether it uses tokens or whatever else is irrelevant.

Also, languages made up of tokens are still languages, in fact most academics would argue all languages are made up of tokens.

Anyway, it's not LLM's that achieve AGI, it's systems built around LLM's that achieved AGI quite some time ago.

Less than 5% of the population knew what it meant to install an app when the iPhone launched. I believe Steve Ballmer ridiculed the idea when asked about it.

A great many amount of people use Android to this day because of its more open nature, and that's despite Google's involvement. If Motorola could go back to its native roots, shake the idea of Chinese influence, and do open source proper, I bet there's a lot more than 5% of the market ready for it.

My company helps companies do migrations using LLM agents and rigid validations, and it is not a surprising goal. Of course most projects are not as clean as a compiler is in terms of their inputs and outputs, but our pitch to customers is that we aim to do bug-for-bug compatible migrations.

Porting a project from PHP7 to PHP8, you'd want the exact same SQL statements to be sent to the server for your test suite, or at least be able to explain the differences. Porting AngularJS to Vue, you'd want the same backend requests, etc..

The Falcon Heavy is $97 million per launch for 64000 kg to LEO, about $1,500 per kg. Starship is gonna be a factor 10 or if you believe Elon a factor 100 cheaper. A single NVidia system is ~140kg. So a single flight can have 350 of them + 14000kg for the system to power it. Right now 97 million to get it into space seems like a weird premium.

Maybe with Starship the premium is less extreme? $10 million per 350 NVidia systems seems already within margins, and $1M would definitely put it in the range of being a rounding error.

But that's only the Elon style "first principles" calculation. When reality hits it's going to be an engineering nightmare on the scale of nuclear power plants. I wouldn't be surprised if they'd spend a billion just figuring out how to get a datacenter operational in space. And you can build a lot of datacenters on earth for a billion.

If you ask me, this is Elon scamming investors for his own personal goals, which is just the principle of having AI be in space. When AI is in space, there's a chance human derived intelligence will survive an extinction event on earth. That's one of the core motivations of Elon.

There was no need for the 15 year old boy who told me this traumatic story of how his sister was killed in his own house to make that story up, because just the fact that they're a liberal family coming from Iran would have been enough information for them to get a visa to stay in The Netherlands based on political persecution.

This happened during Clinton, if you're counting history in US presidencies. And also it doesn't even matter if their sister really was killed. Islamic regimes like the one in Iran are despicable, and would have been even they didn't support goons killing girls for dumb religious regions.

Depends on what you want to hear. The Iranian family in my neighborhood whose father was a doctor fled after Islamist police cut their daughter to pieces in their own home for dressing inappropriately. That's the sort of non headscarf wearing Iranian elite you'll find with an opinion critical of the current regime. I don't know about ostentatious clothing.

To be fair, working on the Rust compiler is also something to do when you're learning Rust. I guess this person is killing two birds with one stone.

If something is literally incredible, then it's prudent to stop and consider whether it should be believed or that you have made an incorrect assumption. In this case, you wrongly assume that Musk is somehow being rewarded for something that happened in the past, or for something that might not even happen. The reality is that the pay package will only have value if Elon manages to dig Tesla out of the hole.

Despite how much conning you believe Musk has done (I won't refute it), Tesla is a company that actually builds cars, and while the Cybertruck flopped and anyone could see that coming from a mile away, that doesn't really affect the Tesla bottom line. That Musk grifted the government into buying them doesn't really do anything besides saving Tesla some money.

I wouldn't buy Tesla shares, I still don't really see their crazy valuation, but I would buy a Tesla car, as they are ostensibly awesome. If you disregard all the lying Musk has done it's still an epic car with unrivaled self driving capabilities.

That he starts talking about something historically has been a sign that some part of it is going to be a reality. You can stand apart from the crazy people who worship the ground he walks on, and still appreciate that he accomplishes great things. Whether it's through conning and grifting, or hard work and keen insight, there are still an electric car company and a rocket company where there weren't before.

Just stop reacting to people believing or shouting things or grotesque behaviors, and just look at the actual reality. It'll do you a lot better than just believing everything Musk says is BS.

The GP was talking about Unreal Engine 5 as if that engine doesn't optimize for low end. That's a wild take, I've been playing Arc Raiders with a group of friends in the past month, and one of them hadn't upgraded their PC in 10 years, and it still ran fine (20+ fps) on their machine. When we grew up it would be absolutely unbelievable that a game would run on a 10 year old machine, let alone at bearable FPS. And the game is even on an off-the-shelf game engine, they possibly don't even employ game engine experts at Embark Studios.

I wonder what's worse, the SFBA-style software development, but also with SFBA-style 2 hour response window to serious bugs like Discord showed, or the old fashioned enterprise report your bug and within 2 months you'll receive an e-mail confirming your report if you're lucky and a letter from a lawyer if you're not.

Yeah it has on occasion, you're right in that it usually doesn't have much of an effect but every once in a while it does. If there's enough of the self-sacrificing users they'll together save a business or a way of doing things. Like running Linux on consumer hardware, or using cash in retail stores.

They don't necessarily have to coordinate, they can use a thousand different linux distro's and literally never talk to each other, and still cause PC manufacturers to keep to a standardized boot process and largely documented hardware so that Linux remains viable.

The author doesn't mean that the technologies weren't inevitable in the absolute sense. They mean that it was not inevitable that anyone should use those technologies. It's not inevitable that they will use Tiktok, and it is not inevitable for anyone, I've never used Tiktok, so the author is right in that regard.

If you disavow short form video as a medium altogether, something I'm strongly considering, then you can. It does mean you have to make sacrifices, for example Youtube doesn't let you disable their short form video feature so it is inevitable for people who choose they don't want to drop Youtube. That is still a choice though, so it is not truly inevitable.

The larger point is that there are always people pushing some sort of future, sketching it as inevitable. But the reality is that there always remains a choice, even if that choice means you have to make sacrifices.

The author is annoyed at people throwing the towel in the ring and declaring AI is inevitable, when the author apparently still sees a path to not tolerating AI. Unfortunately the author doesn't really constructively show that path, so the whole article is basically a luddite complaint.

The rails bugs weren't about Json parsing, they were deserializing into Ruby objects of classes that had side effects, and those side effects led to RCE possibilities. Since those happened, you'll find any deserialization library, especially in dynamic languages, will have a safe (or conversely unsafe) deserialize function to make it more explicit that there's risks involved.

Unsafe deserialization is a very 2010 Ruby on Rails sort of vulnerability. It is strangely interesting that such a vulnerability was introduced so late in the lifetime of these frameworks. It must be a very sneaky vulnerability given how cautious we have become around deserialization since then.

This seems really irrelevant to the browser. I wonder why this was standardized, JavaScript is easily powerful enough to express this. Surely it hasn't been a performance bottleneck for it needing a native api?

I agree with the other two comments, surely almost every frontend webdev has implemented a router in their career unless they never strayed from the major frameworks. It's really not a complicated thing to have to build. I'm not one to look a gift horse in the mouth but I don't see why we're being given this one.

Perhaps, maybe I'm a bad parent but I don't see anything particularly dangerous in Doom 3. It's PEGI 18, but probably fine for teenagers.

Yeah, it's obviously just a crude razor. What I meant is that if it's not on Nintendo, it's a sign you should be more diligent as a parent. Obviously you should always monitor and ideally play the games your kids play yourself, but I can imagine not all parents do that. The Nintendo test is a reasonable alternative, my mom did it in the 90s and I had enough fun games even if I missed out on a bunch of cool PC and Sony titles.

Easy way to know if Roblox is safe for children is to ask the following question: Is Roblox available on Nintendo?

Nintendo would never turn away money, unless they feel it would damage their reputation as a company you can trust your children with.

If they had believed Roblox was safe, they would have 100% taken that money.

I don't believe we need more academic research to achieve AGI. The sort of applications that are solving the recent AGI challenges are just severely resource constrained AGI. The only difference between those systems and human intelligence are resources and incentives.

Not that I believe AGI is the measure of success, there's probably much more efficient ways to achieve company goals than simulating humans.

Sure, if you only consider the model they have no object permanence. However you can just put your model in a loop, and feed the previous frame into the next frame. This is what LLM agent engineers do with their context histories, and it's probably also what the diffusion engineers do with their video models.

Messing with the logic in the loop and combining models has an enormous potential, but it's more engineering than researching, and it's just not the sort of work that LeCun is interested in. I think the conflict lies there, that Facebook is an engineering company, and a possible future of AI lies in AI engineering rather than AI research.

Yeah I think LeCun is underestimating the impact that LLM's and Diffusion models are going to have, even considering the huge impact they're already having. That's no problem as I'm sure whatever LeCun is working on is going to be amazing as well, but an enterprise like Facebook can't have their top researcher work on risky things when there's surefire paths to success still available.

Yeah no, it's just not how it works. They're trying to support fundamental research and they have limited resources to accomplish them. Some random dude who wants to build a company that generates pretty AI pictures is just not the target audience, and he rightly got rejected.

And frankly, the dream scenario that Pieter describes where he somehow would qualify for these resources also wouldn't help kickstart the tech industry, and it's also not how it works in the states.

What does help, and what European governments (at least the one in The Netherlands that Pieter is from) actually do, is more funding for startups. If you're a startup founder in NL almost every angel you talk to has a matched funding deal with the government. That's such a smart way of keeping up with the US. Do you think US startups get free compute from the government? They don't even get subsidies most of the time. What they get is better funding because there's more capital available, and helping investors with that is exactly how you solve that.

Yes, and consider how an animal that lives only 2 years gets Alzheimer's in the first place. They must be genetically engineered to have super turbo Alzheimer's. Normal Alzheimer's doesn't progress so fast I think.

So it seems, but also this seems to have been dragging along since 2023 so it's less sudden than the press makes it seem. The contract with WSS that's underlying the complaint by the former management has apparently been allowed to progress to the point where the US is threatening to simply consider Nexperia a Chinese company for the purposes of sanctions.

Right, I suppose at that point there was some chauvinism or at least fear of being held accountable for it. But the thing that surprised me was that there was actually someone listening at Economic Affairs. Perhaps at that level a board of directors has connections at the government. Either that, or there's some AFM type branch that listens to directors' complaints.

I am sort of surprised we even have an arm of government quick and authoritative enough to be able to intervene here. I always assumed the government would just let any frogs boil.

It's not about non-compete, it's about copyleft. What's blocking redis and and with them a whole generation of open source based startups from embracing true open source licenses is that there is no copyleft license for the cloud age. GPLv3 is 18 years old by now, and the big cloud providers have captured the OSI to prevent a GPLv4 from ever happening.

The open source ethos is not to give code to do as you wish with. It's to ensure users have control over the software and are not controlled by it.

It's not about blocking competition, it's about levelling the playing field. You are supporting the cause of trillion dollar megacorporations in their campaign of capitalizing on open source software over the backs of passionate trailblazers.

The maintainers of Valkey are being tricked by these companies to continue to develop for a project that can be easily exploited.

It makes a difference in that it allows the cloud providers to monetize Redis without having to deal with the Redis team. Which is fair and in the spirit of the original license of Redis, but don't paint it as some great achievement of open source ethos.

What Redis did by embracing a bad license shouldn't be applauded either. But the problem is that there isn't a great copyleft license that prevents embrace, extend & extinguish of great cloud projects. EUPL might've been it, and this weird clause just kills what would have made it perfect.