I was waiting to see when this would turn up on Hacker News. George seems to have a really weird blind spot when it comes to understanding the meta-problem of engineering. When he originally announced this project, he literally explained how terrible the AMD driver is! Yet he spent no time to think why it's like that.
He's shocked and angry that AMD isn't rushing to fix bugs. Well that does sound frustrating, but literally any engineer in the industry could have told you that was going to happen and exactly why. It turns out, everyone else isn't a dumb as you think, and you aren't as smart as you think you are -the problems you're trying to solve are difficult.
It's hilarious hearing him rant though - "I'm not asking for literally anything, except for you to be like holy shit this is the top priority for entire graphics division of AMD". Dude. You're lucky they even speak to you. I can't imagine why Lisa Su the CEO of AMD won't respond to your emails, I mean you represent possibly several thousandths of a percent of increase revenue to them.
AMD is a hardware company. This is how hardware companies behave. It's frustrating, but it absolutely is something that was known before George started this project.
One thing he didn't really pick up on is that the patch that he was angry about has broken English. Why? Because almost certainly AMD outsourced a bunch of their engineering to low cost countries and this is the result. They are not hiring the brightest and the best to work on this. They're selling chips and keeping a good gross margin because that's what the stock market wants.
fjiowf
GeoHot isn't "lucky they even speak to" him, he's spot on about the issues and the utter incompetence of most middle management and what needs to be done if they want to compete in the GPU space.
And it's not because the problem is "that hard", it's because AMD doesn't have many skilled engineers like you mentioned at the end of your post (basically contradicting yourself!), and won't prioritize the right issues.
rchaud
Maybe George can volunteer as an unpaid intern, like he did for Twitter. He certainly solved their problems! /s
SilverBirch
Do you understand how these companies resource their customer service divisions? Because I do. They most likely have zero direct customer service for end customer support, you'd need to be purchasing >$100m of hardware to be getting direct engineering support. It's more likely that the person supporting him internally is getting funded by the Sales & Marketing budget than by support.
All George is really asking for is AMD to entirely reorient their business, change their sales and marketing strategy, hire in a cadre of extremely expensive software engineers and give up the entire margin they make on the chips. No big deal.
It's like... sure, it'd be great if they were suddenly a fantastic new company that operates differently, but we live in the real world here.
omeze
I think there’s an element here thats missing: companies can fund small tiger teams (a la skunkworks) if theres executive buy-in to solve specific targeted problems. A small elite unit of engineers with executive support and external partners can certainly make a difference, without reworking the entire company. I think geohotz has the expectation that AMD would try such a thing, as Ive seen and heard of it many times in industry (software or hardware).
SilverBirch
Tiger teams can work in some scenarios, but it seems unlikely that's going to work at AMD. The reason the driver is bad is a lot to do with the fact that the driver team have to basically make lemonade out of the lemons that the hardware team hand them. The hardware team have a difficult enough job just to get the hardware to work and they're where the core success of the business (traditionally) was seen. So the driver team have to do the best they can, but at the end of the day they can't independently solve the problem because they're downstream of the hardware. So yes, you could spend some money and get better engineering on the driver side, but there's a limit to how much you can do without re-prioritizing the hardware team to be more responsive to software requirements and at that point you're talking about full scale organisational change. Only a handful of companies can do that. Look at Gelsinger trying to change Intel - the stock market is not thanking him for his efforts, we're at the point where they're a legitimate acquisition target for Nvidia (ignoring anti-trust concerns)
red-iron-pine
* elite engineers are expensive
* tiger teams or specialist troubleshooters and project pushers are almost certainly already a thing in AMD
* their "this shit is on fire assign it to a tiger team" criteria may have absolutely nothing to do with Geohot's complaints. things like legal or compliance issues, Taiwan getting attacked by China, high rates of turnover on [specific team], or a problematic design may rank a lot higher than some rando who got headlines a few years back for some good cracks.
Keywords were enough to kill my interest within seconds.
Maybe is how americans talk on a day to day basis. Not being with them I wouldn’t know. But I regret wasting few seconds of my time on this “shit” already.
someguydave
Nvidia makes hardware and their driver does not crash
kramerger
First twitter now AMD?
Come on George, if you continue giving up this easily someone may make you a product manager at Google soon
flipgimble
tinygrad and Hotz came on my radar only recently, so I looked into it. the most insightful comment, thats proven even more accurate with time, is this one:
https://www.hackerneue.com/item?id=36066136
quote: "he gets excited about the currently hot thing in tech, makes his own knockoff version, generates a ton of buzz in the tech press for it, and then it fizzles out because he doesn't have the resources or attention span to actually make something at that scale"
aigoochamna
A close friend of mine went to highschool with George. Everything I read about him coincides perfectly with all of the highschool stories. At least he has stuck true to himself all these years :p
He's shocked and angry that AMD isn't rushing to fix bugs. Well that does sound frustrating, but literally any engineer in the industry could have told you that was going to happen and exactly why. It turns out, everyone else isn't a dumb as you think, and you aren't as smart as you think you are -the problems you're trying to solve are difficult.
It's hilarious hearing him rant though - "I'm not asking for literally anything, except for you to be like holy shit this is the top priority for entire graphics division of AMD". Dude. You're lucky they even speak to you. I can't imagine why Lisa Su the CEO of AMD won't respond to your emails, I mean you represent possibly several thousandths of a percent of increase revenue to them.
AMD is a hardware company. This is how hardware companies behave. It's frustrating, but it absolutely is something that was known before George started this project.
One thing he didn't really pick up on is that the patch that he was angry about has broken English. Why? Because almost certainly AMD outsourced a bunch of their engineering to low cost countries and this is the result. They are not hiring the brightest and the best to work on this. They're selling chips and keeping a good gross margin because that's what the stock market wants.
And it's not because the problem is "that hard", it's because AMD doesn't have many skilled engineers like you mentioned at the end of your post (basically contradicting yourself!), and won't prioritize the right issues.
All George is really asking for is AMD to entirely reorient their business, change their sales and marketing strategy, hire in a cadre of extremely expensive software engineers and give up the entire margin they make on the chips. No big deal.
It's like... sure, it'd be great if they were suddenly a fantastic new company that operates differently, but we live in the real world here.
* tiger teams or specialist troubleshooters and project pushers are almost certainly already a thing in AMD
* their "this shit is on fire assign it to a tiger team" criteria may have absolutely nothing to do with Geohot's complaints. things like legal or compliance issues, Taiwan getting attacked by China, high rates of turnover on [specific team], or a problematic design may rank a lot higher than some rando who got headlines a few years back for some good cracks.
Keywords were enough to kill my interest within seconds.
Maybe is how americans talk on a day to day basis. Not being with them I wouldn’t know. But I regret wasting few seconds of my time on this “shit” already.
Come on George, if you continue giving up this easily someone may make you a product manager at Google soon
quote: "he gets excited about the currently hot thing in tech, makes his own knockoff version, generates a ton of buzz in the tech press for it, and then it fizzles out because he doesn't have the resources or attention span to actually make something at that scale"