Hacker Newsnew | past | comments | ask | show | jobs | submit | cakealert's commentslogin

> No, based on cURL's history, it really seems like they would love to have found a really novel bug.

You just confirmed that you didn't read the article.

"Eventually, I was instead offered that someone else, who has access to the model, could run a scan and analysis on curl for me using Mythos and send me a report."


I'm not sure how that proves I didn't read the article ?

Someone external to the curl team ran the test. If that third party found a severe CVE that they could use across all the global curl attack surface, and did not disclose it back to the curl team, the third party could keep using the exploit until discovered independently.

Why the focus on Lynn?

Focusing on his critics is more illuminating and damaging to your presumed position:

Systematic review by Wicherts et al: "In light of all the available IQ data of over 37,000 African testtakers, only the use of unsystematic methods to exclude the vast majority of data could result in a mean IQ close to 70. On the basis of sound methods, the average IQ remains close to 80."

They of course follow it with the conjecture: "Although this mean IQ is clearly lower than 100, we view it as unsurprising in light of the potential of the Flynn effect in Africa (Wicherts, Borsboom, & Dolan, 2010) and common psychometric problems associated with the use of western IQ tests among Africans."

It's always curious how "common psychometric problems associated with the use of western IQ tests among Africans" don't carry over to economics and other things. Wouldn't you expect them to have similar problems with other western "ideas"? Also interesting how Easterners adapted to western IQ tests so well they are better at them than the West.


Part of the point of the Wicherts papers was to refute the Lynn data, observing that even taken on its own premises Lynn's team pretty clearly excluded data unfavorable to the conclusion they wanted to draw. But look at Wicherts 2009, at the samples they're talking about. One of the largest was 800 students in Nigerian high schools (a test arranged by IQ researchers to for a cross-cultural comparison, back in 1981). Lynn's data includes, and is materially influenced by, a sample of 59 Senagalese children who were tested while recovering from malaria. The malaria thing is just funny, but it's the numbers that stick out to me.

I'm not getting further into the details here because the easiest-to-understand point here is that there are not in fact programs to generate reliable "average IQ" numbers in different countries. I am struck by the fact that message board nerds from America believe these programs exist, when almost none of us have ever taken an IQ test.


The purpose of IQ tests is to derive a value that would predict other values. The inverse is also true.

Given a multitude of such values it would even be possible to get back to a precise IQ value.

IQ tests are just factor analysis artifacts. You can dream up 100 questions that you conjecture may have something to do with intelligence and not even know the answers and have 10,000 people answer them. Product of the factor analysis of the answers will yield a normal distribution which you can then center on 100. Calibration can be more complicated than that but you get the point.

The natural circumstances people find themselves in are also just noisy questions.


This argument is basically a rejection of the science of IQ testing and I am here for it.

This is fairly naive, Elon isn't the only investor in SpaceX.

My guess is "that they did the math" and had an engineering study which convinced them that getting AI datacenters into space will make sense.

It's also not hard to imagine why, the process alone once perfected could be reused for asteroid mining for example, then mirogravity manufacturing, either of which alone would be enormous capital intensive projects. Even if AI dataenters in space are break-even it would be a massive win for SpaceX and leave their competition far behind.


Are you a bot or are you just stupid?


There are several other companies that have announced efforts to try data centers in space.


I know this is hackernews and we like to get hyped up for new technologies, but, like, this just straight up isn't happening.

There is no benefit to putting data centers in space versus the giant cost that you would incur by doing so.

Can people please try and use their fucking brains for a second?


> Can people please try and use their fucking brains for a second?

Have you considered that people smarter than you think it is plausible?


> Have you considered that people smarter than you think it is plausible?

I know many people smarter than me, plenty of them who have spent careers building data centers, and not one of them think this is plausible.

You should consider whether people smarter than the average investor are pulling a fast one.


Maybe we are talking about different things here?

I don't doubt spacex can fail at this.

I also don't doubt we are fairly close to making this plausible.

> plenty of them who have spent careers building data centers

Famously, plenty of people who have spent careers building rockets would swear that reusable rockets would absolutely never work.


>I also don't doubt we are fairly close to making this plausible.

Maybe you should doubt that. There's literally no reason to think this is plausible besides some hype merchants' say-so.


> some hype merchants

Excluding Spacex:

Nvidia, Google, China, European Commission, Blue Origin

And this being HN, a YC funded company has put a single GPU rack in space and demonstrated training a reasonable sized model on it.

But yeah, it's all hype, sure.


On the off chance you're sincere and not just heavily over indexed into Elon stocks:

It's trivial to understand why this is all hype if you pay attention to physics, as another commenter suggested earlier.

https://en.wikipedia.org/wiki/Stefan%E2%80%93Boltzmann_law

Assume you're radiating away the heat for a single B200 (~1kW), and the max radiator temp is 100C, you find A = ~3m^2.

So that's 3 square meters per GPU. Now if you take into account that the largest planar structure deployed into space is ~3k m^2 (https://investors.lockheedmartin.com/news-releases/news-rele...), you're looking at 1000 GPUs.

That's a single aisle in a terrestrial data center.

Cost to deploy on earth vs satellite is left as an exercise to the reader.


You are missing one important thing here.

You do not radiate all the heat away from a GPU, a modern GPU can run pretty hot. Also look up how this is getting better for the next generation of GPUs.

Maybe repeat your calculation with updated assumptions?

But even if you were completely right, your argument is that we can't do this tomorrow, yes I agree. Typical technology development cycles are about 5-10 years.


> You do not radiate all the heat away from a GPU, a modern GPU can run pretty hot.

Fascinating. Tell me more.

Where does the heat energy that isn't radiated away go?


>You do not radiate all the heat away from a GPU, a modern GPU can run pretty hot.

LOL. If you don't radiate the heat the spacecraft just gets indefinitely hotter (until it glows and the heat is forcibly irradiated). It's space, there's no fluid to provide convection.


Have you considered that people smarter than you are scamming you?


Yep, definitely being scammed by not dismissing things outside my area of expertise out of hand.

I wish I had your confidence about everything!


I am yet to see any actual numbers showing how the economics of this would work or compare to the cost of building traditional data centers.

Please come back to reality.


> Yep, definitely being scammed by not dismissing things outside my area of expertise out of hand. I wish I had your confidence about everything!

Instead you put your confidence in Elon, who has zero expertise in this area?


> Instead you put your confidence in Elon

No, I put confidence my ability to do a web search, pretty rare skill nowadays ;)

You'll see that none of these are Elon/spacex, hopefully?

https://medium.com/@cognidownunder/google-just-announced-the...

https://www.ycombinator.com/companies/starcloud

https://www.informationweek.com/it-infrastructure/lunar-data...

https://ascend-horizon.eu/

https://www.axiomspace.com/orbital-data-center


I often wonder that same thing about many hn commenters!


Yes that guy is extremely weird, he should delegate operations and community management to someone who isn't weird and stick to development.


The same way that having motorized farming equipment was a race to the bottom for farmers? Perhaps. Turned out to be a good outcome for most involved.

Just like farmers who couldn't cope with the additional leverage their equipment provided them, devs who can't leverage this technology will have to "go to the cities".


Please do read up on how farmers are doing with this race to the bottom (it hasn't been pretty). Mega farms are a thing because small farms simply can't compete. Small farmers have gone broke. The parent comment is trying to highlight this.

If LLM's turn out the way C-Suite hopes. Let me tell you, you will be in a world of pain. Most of you won't be using LLM's to create your own businesses.


But modern tillage/petrol based farming is an unsustainable aberration. Maybe a good example for this discussion, but in the opposite direction if it is.


This wasn't even true 10 years ago...


The idea that a group of people would spend so much of their time trying to get linux to work on Apple hardware through reverse engineering always seemed absolutely crazy to me. I would never consider buying Apple hardware precisely because it doesn't support linux and the work they put in achieves nothing because the risk will always remain that they will lock the hardware further. Nevermind the fact that they will likely never fully reverse engineer all the components.

It just seems like a completely pointless endeavor... perhaps some people buy into it? why would anyone buy overpriced hardware with partial support that may one day be gone? the enhanced battery life doesn't really hold much appeal to me, and the arm architecture if anything is just another signal to stay away.

The only thing that makes sense to me is that they wanted the achievement on their resume, and in that given recent developments they succeeded?


You overlooked the UTM app on the App Store (and open source available too), which wraps Apple Silicon virtualization excellently, or you can use Qemu (which I don't).

I used to use Asahi, but the sleep modes power drain was tedious.

With UTM, I install a latest Fedora ISO (declaring it a "Linux", which exposes the option to skip QEMU and use native Apple Silicon virtualization.

It's fantastic. I mention this only because it's been super useful, way better than Asahi, with minimal effort.


The hardware isn’t overpriced, it’s best in class. It’s just that that class isn’t what you’re looking for, and as a Linux user it’s not for you, which is valid! But the hardware for what it is is one of the absolute best price to performance ratios on the market right now and I’m tired of people pretending it isn’t. You can get a brand new m4 MacBook Air for under $800 right now, and that’s simply one of the best deals around. For an M2 for asahi Linux? Second hand the prices are even better.


It's like Hackintosh all over again but with Apple hardware rather than their cursed software.


Maybe they just needed a hobby. I for one think it's a pretty cool one.


A sufficiently seeded torrent is a high latency static CDN.

You just need a client that can make use of it.

I'm not sure if anyone will be interested in making one however, you can already get a patched Spotify APK from the usual mobile piracy spaces that's good enough.


Wasn't popcorn-time basically video streaming backed by torrent ? Why can't it be the same for audio ?

The metadata is 200 GB which can be easily indexed and could be made searchable, then you download only what you need


Now that's a cool idea.


When they say "worse" they do mean the AI will get better which will be worse because they are ideologically opposed to AI.


I'm not ideologically opposed to AI. The problem will get worse because while the quality of the music will improve, it will still be bad and there will also be a lot more of it.

We aren't really short on music. Diluting the good stuff with 100x more mediocre filler is not a good thing.

If AI generated music ever actually becomes good then that's another story but that is quite a way off.


when you combine a residential proxy with a tool like curl-impersonate (there are libraries in Go for this type of fingerprint spoofing now) they dont even show up as scrapers anymore, just users. especially when they adjust timings to mimic humans.

clouflare only blocks the most dumb of bots, there are still a lot of them.

this is why cloudflare will issue javascript challenges to you even when you are using google chrome with a VPN, they are desperate to appear to be doing something. and every VPN is used to crawl as well. a slightly more sophisticated bot passes the cloudflare javascript challenge as well, there really is nothing they can do to win here.

i know some teams that got annoyed with residential proxies (they are usually sold as socks5 but can be buggy and low bandwidth) so they invested into defeating the cloudflare javascript challenge and now crawl using 1000's of VPN endpoints at over 100 Gbit/s.


Is "residential proxy" another name for an hacked/owned computer that the bots have access to? Or are there legitimate services that sell access to residential IPs?


People legitimately sell egress. It's "free" money. But of course, if you have a botnet, you can sell that through the same channels, no one is looking too closely.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: