djcapelis

If this is what failures from a “dead” project looks like we’re only a few more failures away from some working rail: https://youtu.be/sa-X1qqxvJU

(I’m much more aware of the California HSR project than most people on the Internet.)

I feel like your comment is spot on for the usual Texas vs California dialog:

High speed rail in Texas: 0 miles built, zero miles under construction, projects stuck in regulation waiting to begin work[1]

High speed rail in California: Active construction, most ambitious real HSR project in the country, project delay by regulatory hurdles… mostly because conservatives tried to sue it to death and failed.

Texans on their progress vs California: California is a failure.

This whole meme is because a couple struggling Silicon Valley companies who stopped being able to make it in SV decided to relocate to Texas to go out to pasture along with a couple rich people. Y’all think that being a launch site for SpaceX, which remains a company with a California headquarters, and a majority California positioned engineering team is hot stuff. Y’all celebrate how many tech jobs come to Texas when California companies open branch offices for support and secondary roles in your state and crow about how you one upped the Californians as you landed those amazing customer support roles by selling corporations pieces of Texans and the time of your citizens for cheaper. Then y’all lecture us about how you just got HP’s enterprise spinoff, when we have the actual HP… y’all need to stop claiming to be doing better at areas you’re doing worse on if you want to figure out how to actually do better.

[1] https://www.texastribune.org/2020/11/17/houston-dallas-bulle...

Why don’t you look through these job listings and tell me which state does most of the science on those rockets: https://www.spacex.com/careers/?department=

I don’t think SpaceX is a great example of “Texas science” given the vast majority of their major science and engineering work is located in their California offices.

> Why should outdoor dining be closed, yet indoor retail stores still be open?

That’s easy. They shouldn’t.

And yes, it is that simple. California is within days of running out of ICU capacity in whole areas of the state just given our current case progression rate even if you immediately found a way to freeze all transmission tomorrow.

Additional human contact directly costs even more lives right now, in this specific time.

This is exactly how the hardware works and what micro-ops are, on any system with a u-op cache or trace cache those decoded instructions are cached and used instead of decoding again. Unfortunately you still have to decode the instructions at least once first and that bottleneck is the one being discussed here. This is all transparent to the OS and not visible outside a low level instruction cache though, which means you don’t need a major OS change, but arguably if you were willing to take that hit you could go further here.

You are correct that with a lot of effort you can get a compiler to emit VLIW instructions that can claw back some of these problems and have some, but not all, of the information you have available to the processor core at runtime which can make things slightly smarter. And with all of that effort which involves changing software distribution models, jettisoning energy efficient flexible heterogenous architectures, and abandoning SMT, you might be able to get something close to performing as well as… what we already have today without VLIW.

Not sure why anyone thinks that’s an argument for doing it.

I think you are correct that there is a section of the field enchanted by the idea of VLIW and desperately hoping to justify any possible reason why building these chips was less of a bad notion than they’ve turned out to be each and every time.

I don’t think we’re really learning much by pretending building a new one is a good idea. Or that the last efforts had much value.

No one is saying don’t make progress on compilers. Look around you and you’ll see the things you just referenced already exist in practice in production today. (Among others: compiler/architecture co-design is a well established and fertile area of work. Compiling to intermediate formats which specialize for a device later is in use to some degree in App Store, which keeps programs in an intermediate format and thins binaries down to specific slices on demand for iOS devices, etc.) None of these technologies justify building a new VLIW chip or have frankly much to do with those efforts. The best thing we could learn from the huge amount of effort that’s gone into VLIW architectures is actually learn the lesson that in front of us, demonstrated repeatedly, that unless something substantial changes, VLIW does not produce a better architecture for processor cores and the assumption that it would was wrong.

> Compilers "failed" because VLIW is fundamentally not a particularly useful idea.

I think the problem isn’t just that it’s not particularly useful, but that it’s an actively bad idea. It statically encodes ILP that we get dynamically in chip designs today, which means that ILP can no longer react to changing conditions, like literally any core architecture changes as the underlying hardware evolves (or even on the same die, say finding itself running on an efficiency core rather than a performance core or vice versa) or running in an SMT environment in which it is not the only instruction stream going through the pipeline.

> we're also going to see someone give VLIW another shot

Dunno why people would keep thinking this is a good idea, but if someone wants to try eating it with a new VLIW design yet again, I’m always down for some popcorn.

Dynamically finding ILP at instruction dispatch time is always going to be smarter than a compiler trying to guess at compile time. IMO, it’s worth the area. Even if you assume the compiler gets it right and finds tons of ILP, it’ll get it wrong for the next hardware rev when your core inevitably changes and you have to break apart your VLIW instructions and get ILP back dynamically anyway. To say nothing of shared pipelines in an SMT processor, where VLIW basically runs counter to SMT and the only way to know which units are going to be available on a specific cycle is during runtime.

> and they're going to succeed this time.

I don’t see any reason why VLIW would succeed more in future attempts than in past attempts.

I think it’s a good point actually. I think it’s reasonable to say swift isn’t intended as a language to experiment with a specific concept, it’s a six year old production language designed for a general set of usecases where it’s good at a bunch of things that the languages it is replacing weren’t as good at. There’s some interesting discussions to be had about tradeoffs it makes but a question like “what’s the one thing swift does that other languages don’t do” is not the right way to evaluate a production language. It might explain why some people don’t find it interesting but swift was built to be a better tool, not merely an interesting one.

A lot of really good programming languages are “uninteresting” in this way, and that’s probably a good sign, not a bad one.

Are they? If so, we’ve been saying them for about the last 30 years and things seem to be fine.

Pick your favorite ranking of world universities and let me know how many California universities you have to pass before you find a Texas institution on the list.

It’s nice that you’re so invested in your fever dream of CA’s downfall. I hope where ever it is you live turns out nice for you. As long as we don’t descend to full on fascism next year I think I’ll buy a house here. I like California. I stay here because of it. Maybe if enough people who think that Texas is a good idea leave people then who want to be in California can replace them and we can build more housing stock together! :)

FWIW, I’m pretty much entirely for bulldozing local zoning law and control over a lot of housing and have been delighted by our recent progress there and often vote accordingly. I think this is our major issue, not asking people and companies to pay for advancing our communities.

> If you make it too costly to do business, business will leave.

They’ve been saying that for 30 years and the numbers have not significantly changed.

> That or a company from somewhere else will use less taxes and regulations as a competitive advantage. > They'll even take losses like shutting down factories.

Companies based here are some of the most profitable in the world. Your statements are pure fantasy unanchored to any facts.

> Cheaper housing, less homelessness, less expenses.

If as many people wanted to live in California as want to live in Texas, it would be cheaper to live here too.

In addition, the average California household also makes more, by a not insignificant percentage.

I live in CA rather than Texas because I expect the state to take some of the wealth we create with the common infrastructure we’ve built like excellent public education etc and reinvest it in our communities.

That’s why so many companies chose to be here and have their highest paid positions here too.

Other communities in other states are welcome to bend over backwards to enthusiastically attract our tablescraps and feel like it’s a win to their communities to have them. But they don’t get to expect to be taken seriously if they tell us to change how we run our state to be more like them.

Wake me up when Tesla moves their management and engineering out of CA rather than just writing snotty tweets on the Internet.

Feel free to scroll through yourself: https://www.tesla.com/careers/search#/?country=-1&city=-1

Not only are most of the US based jobs on Tesla’s site still located in CA, most of the better paid higher qualification jobs are too.

(Their headquarters remains in CA too.)

This nonsense pundit piece is talking about ~700 or so commercial entities leaving CA. This is a state where that number of corporate entities is a rounding error. I’m shocked the number isn’t higher. We are a long way from Detroit and have an industrial base more diverse than a small handful of powerhouse companies in a specific sector. We also have invested in creating infrastructure for innovation to a level where it isn’t an accident when things happen here.

I think using Detroit as a scare tactic to encourage desperate policies to retain companies in a state that’s lousy with them is a pretty far cry from reality.

> It seems like people in Texas on average are doing better than California.

How do you figure?

Companies are plentiful and as the fifth largest economy in the world we don’t have to fight like pigeons over our cast offs like other states do. In the time that ~700 commercial entities left, how many more do you think were created here?

Housing here is a legitimate problem but the companies that flee to Texas because of lower taxes aren’t worth keeping.

What did you expect? The Hoover institution is a fundamentalist think tank and their analysis is written entirely to pitch lower taxes and less regulation as always being the right thing. In short, they are living up to the level of brilliance of the economic policies of their founder.

I always find it amusing when a policy thinktank founded by Herbert Hoover wants to dispense advice on how to build a strong economy.

Welcome to California. While we have plenty of problems, fundamentally, if your company can’t cut it here and decides to relocate to Texas, that’s not an indictment of California. It’s weak companies fleeing to a place where the government is more willing to subsidize their operations at the expense of their residents.

US society, discourse, media, and government structures have sidelined experts to the extent we’ve finally managed to forget they exist. Instead we’re left with reporters who write articles that bemoan that since they’re lost and unable to determine truth in a sea of technology, we all must be.

So... no?

> It's the #1 way to pay progression, sadly

This has not been my experience.

Yeah I agree with you. My post was written in haste and yours captures the intent much better. Thanks!

Well first off I think any company using leetcode as a screen for their candidates is out of the minds. (It seems to be a popular way to be out of their mind but that doesn’t make it a good idea. :))

But implementing a trivial unoptimized solution for the problem you linked is feasible enough in 45 minutes. Talking idly through how one might tweak the algorithm to optimize, or adopting an optimized algorithm if one occurs to you as you’re implementing the basic one also seems fine. If they want a tested, optimized, and implemented novel solution in that period, they are testing for people who grind leetcode and not people who are good engineers. I would have likely saved myself the time by not applying to a company like that in the first place!

Screens should generally focus on finding candidates worth talking more with. Leetcode exercises are low value for that.

If you want to learn anything, don’t work there. Work at a place that selects for better talent than mindlessly picking leetcode exercises as a screener. You will learn more, you will be at a company that does better, etc.

(I have never screened a candidate using leetcode. I have never been screened using leetcode. The one time someone tried to screen me using stupid questions like this I finished our session and promptly terminated the process when they called me back for a future round because while I may have passed their screen they emphatically did not pass my screen. I don’t study for interviews, nor do I expect people who I interview to study. Nothing I’m going to ask is something you can cram about the night before. This isn’t a college midterm.)

> all these jobs were in the last 7 years.

Question here... I hope I’m not being a huge jerk here, but uh. Does this feel like this strategy is working well for you?

Did you like any of these positions?

60 interviews in 30 days is not a process in which you’re carefully learning and calibrating to tune yourself, extracting lessons from each one.

60 interviews in 30 days is throwing an entire pot of spaghetti against a wall and not even waiting to figure out if you cooked it right before dumping another pack on the stove and firing back up the burner.

This not a person carefully working towards a thing or getting practice to calibrate. This is a person who has read too many blog posts like this and regrettably and agonizingly thinks the process of finding a job involves interviews in bulk and funnels and a spreadsheet at the end of which pops out some offers that are then mostly weighed based on numbers and one gets accepted.

This process is what way too many people who blog about their job searches do and it is completely unhinged.

And it is also, not what most of the good engineers I know actually do.

Don’t interview in bulk! Look for the positions you really want and think can grow you in a way that matters. Then find out what skills you need to do that. If you have them, go talk to that set of organizations and tell them about your skills! People desperately need good people!

If you don’t have those skills yet, either look for a different match... or think about where you’d need to go to get those skills. ... and then go talk to those organizations!

Figure out something you’d like to work on. Figure out where your skills are the best match for those things. Then interview for those!

Your conversations will be more focused. Your approach will be more focused. You will get jerked around less. And people will be thrilled to talk with you and figure out how your passion aligns with what they need. And if it’s a match, they will want to hire you.

Do not mindlessly grind. Do not scattershot. Do not play games.

All of that is unfocused, unsuccessful and unlikely to connect you with a position you even care about!

What ends up happening with these people is they set up this huge funnel and then at the end of it they play offers against each other and have agonizing decisions about which numbers to accept for a set of positions they aren’t even excited about instead of focusing on what matters:

1) Which position will grow you the most?

2) Which position do you think you can exceed expectations in?

You want to optimize for the sweetest mix of both of these. And both of these factors are complementary to each other. Organizations grow the people exceeding expectations and you will exceed expectations most in a position that provides you the glide slope of growth which is the best match for your interests.

If you’re early career, an offer is the start of a business relationship. Not the destination. And if you’re doing it right, the initial offer is what they want to pay you to try things out. What they will want to pay to keep you happy will be more than that after you join. If you’re later in your career maybe you and the company are so correctly calibrated to each other and in sync that your initial offer hits it exactly on the head and the adjustments are for growth from there. Whatever. Who cares. Even if what you care about is money: go for the job where you will grow the most and provide the most value, because ultimately that’s where you’ll have the most leverage to grow numbers either in the same company, or in a different one.

To say nothing about the obvious fact that this is a huge chunk of your time each day you engage in it and it’s 100% worth it to do something you’re excited about instead of just taking the offer for the person who pays the most. This field pays well. A 150k/yr base isn’t a disaster even if you could have gotten 220k/yr elsewhere. Which is totally a huge difference! But will round away over time when your growth is explosive and the engineer plugging away logging hours at a 220 base goes nowhere, stalls and ends up with uninspiring refreshes and trajectory.

Routinely failing a simple tech screen is not a symptom of stretching. It’s a symptom of probably not being in a place to be blogging interview lessons for others. Totally credit to this dude for putting himself out there, but I think he’s radically miscalibrated to be doing so in a way where he’s had this experience and is off handing out advice. To extend a more accessible analogy, in universities, the people who pass the course the first time are picked as tutors. Not the people who re-took it seven times because it was a stretch for them.

> it might be because they're trying to do harder things

I literally work on one of the, if not the, top team for what I do in the entire industry.

I’ve failed plenty of times.

I’ve never failed a basic tech screen though...

They’re not supposed to be hard!

Hotter tip: toss the advice from people who interview a lot and aggressively and seek advice from the people who wander into opportunities seemingly without effort. Ask them how they approach their job search.

Most of this advice is bad or fixes problems that come from scattershooting an interview process by bulk for nonsense you don’t even care about.

This blog post outlines the most miserable way to do a job search. And the way to get rejected a lot.

> Those are those sort of numbers I'd expect for every candidate.

> The point of a technical screen is to filter people out.

I mean sure. But I’ve never once been rejected on a tech screen. Most people I’d want to take advice from don’t get screened out often either.

So much interview advice is just awful and these “grind away mindlessly” narratives are actively nonsense that does not correlate to getting a well matched and awesome position.