wegs

52 ms·

391.
by wegs 6y ago
Maybe. The other possibility -- with people 3-4 feet apart -- is that demand will outstrip supply. You'll have far less capacity for people to eat, perhaps by much more than fewer diners. Two things are sure: (1) Things are gonna get w
392.
by wegs 6y ago
I'm keeping my money there since it's backed by the US economy. Yes, the value of the economy is lower than 6 months ago, and while likely be lower in a year. What else should it be backed with? Dollars are subject to (hyper)infla
393.
by wegs 6y ago
Please show me a business (aside from Amazon, obviously) who needs Alexa. Most businesses need a word processor, a spreadsheet, and some kind of database for managing employees and inventory. A 486 does that just fine. Most businesses deriv
394.
by wegs 6y ago
Awesome! I did not know that.
395.
by wegs 6y ago
shrug I'll bet my business on waiting an extra 15 minutes for analytics code to run. Seriously. There's little most businesses really needs that I couldn't do on a nice 486 running at 33MHz. Now, if a $5000 workstations g
396.
by wegs 6y ago
It's not a safe bet. I've build project under AGPL, and made plenty of money doing it. There are places where open is good business, and there are places where proprietary is good business, and there's everything in between.
397.
by wegs 6y ago
You're missing the point, and taking everything as an attack on NVidia. It's not an attack on NVidia, or have anything to do with Nvidia versus Khronos. It's clear you've got enough baggage there that I won't go the
398.
by wegs 6y ago
Well, there's a question of whether the GPU has to be fed fast enough. My CPU is nominally rated at around 150-200 gigaflops. My GPU is rated at about 5 teraflops. That's about a 30x speed difference (and there are obviously faste
399.
by wegs 6y ago
On-boarding mostly happens before users even install it: 1) Run the page by users unfamiliar with it, and see if they'd install it, why, or why not. Especially language learners (e.g. go into a Chinese classroom). DON'T tell them
400.
by wegs 6y ago
Aside from Android support, my key suggestion would be to clearly think through on-boarding. Visiting the page, I see a $13 monthly plan, in an era of free and $5 iPhone apps. I don't know if it's any better or worse than free, so
401.
by wegs 6y ago
It's not about blame. It's about getting to an ecosystem where GPGPU is used for things beyond deep learning, bitcoin mining, video encoding, and similar niche applications to one where I can fluidly use MIMD to speed up my JavaSc
402.
by wegs 6y ago
Yes. To have openness work, you need to execute well too.
403.
by wegs 6y ago
> There is also the problem of needing huge amounts of capital, where most of this work can only be done by exceedingly well funded groups with deep ties to hardware divisions in question. The future of hardware innovation comes from bil
404.
by wegs 6y ago
Yes. I won't bet my business on a one-vendor solution with a medium-sized community which might disappear at some point. If CuPy supported NVidia and AMD, and was folded into Numpy, I'd buy the biggest, beefiest GPU I could find o
405.
by wegs 6y ago
I don't underestimate the complexity. But I do claim that the complexity can and should be hidden behind programming language constructs. I've worked both on the design of MIMD hardware, back when I was a graduate student, and on
406.
by wegs 6y ago
You're making a lot of lousy assumptions. As a few points of reference: * My list comprehensions run over gigabytes of data (but sometimes 3 orders of magnitude bigger or smaller). Stream processing of big data. It's not deep lear
407.
by wegs 6y ago
Intel was forced to license to AMD for government contracts. There's a super-complex story there I won't get into. There were a few clone vendors aside from AMD. None were ever a serious threat, and AMD itself didn't become m
408.
by wegs 6y ago
It's irrelevant to researchers. Research operates on rapid cycles: prototype, publish, move on. It does impact businesses. It doesn't prevent adoption for e.g. deep learning, but I haven't seen e.g. GPU-based databases reach
409.
by wegs 6y ago
Yeah. That's been my general problem with adopting NVidia for anything. They make good hardware, but there's a lot of lock-in, and not a lot of transparency. That introduces business risk. I'm not in a position where I need
410.
by wegs 6y ago
The debt forgiveness can happen here too. That's the critical piece now. A month ago, a debt freeze would have done it.
411.
by wegs 6y ago
Well, the goal of an economy is to (1) Keep people working productively (2) To distributes the fruits of production equitably. What constitutes productive work is disputed, of course. Most people agree growing food and teaching kids is prod
412.
by wegs 6y ago
I just want to mention this is why I believe Google will never be able to compete with AWS, or otherwise be credible in the B2B space. You're relying on automated systems which can take down your business on a whim, with no recourse. W
413.
by wegs 6y ago
Everyone knows what happened. But the result was that a huge part of our history was lost. When we write a book, it is archived in the Library of Congress. As publishing moves online, and we deprecate technologies due to technical convenien
414.
by wegs 6y ago
I'm arguing: (1) The poster arguing this was 10 minutes was being dishonest. (2) As the original article said, it's a different internet if the only way to get something up is to go through Square Space or Github pages. It's
415.
by wegs 6y ago
Ask your mom to "put an old HTTP server behind a reverse proxied nginx instance with auto-renewing letsencrypt certs" and time her. See if it takes 10 minutes. Then ask someone to foot the server bill. See how many takers you get.
416.
by wegs 6y ago
Making the computations type-level correct would be really neat! That'd be a huge improvement! I don't think the language needs to be statically-typed to do that. I could easily extend Scheme or Python to support that. All you nee
417.
by wegs 6y ago
The article quotes 56%, as the negotiated rate on NiH projects, not 54%, but the rate varies by funder and by project. The 2/3 number is from an actual project I had insight into as the base overhead rate. What's a little bit hidd
418.
by wegs 6y ago
Most money which comes into MIT passes through overhead. That means if a foundation donates to MIT, a bit over 1/3 of that money might ends up with whatever they donated to. A bit under 2/3 might go into the general budget (overhe
419.
by wegs 6y ago
This is still true. Your second sentence is correct, but the rest is not. Most research does come from within the system. Unconventional research does not and cannot come from within the system. Funding comes from calls from proposals which
420.
by wegs 6y ago
I think so too. My concern is about Tesla's leadership: 1) Elon Musk has made many comments downplaying COVID19, and 2) In photos (even in their Tesla Return-to-Work Playbook), the Leadership Team doesn't wear face masks (only the

More ›