1,305 karma · joined July 2, 2019
Should I try it again?
I also might be confused, but I thought they kept some key parts proprietary. I'm now looking, and that doesn't seem to be the case. Perhaps something changed (perhaps even my memory!). I'm a lot more happy to pay for 100% open-source than 90% open-source. Is it fully open-source?
I'd love to pay for a reliable, hosted, open-source solution.
It's not at all uncommon.
I've done business with Google multiple times in my career, and I could name a dozen such incidents. They're just not public.
For the most part, there wasn't anything the user did; just Google algorithm bugs. In one case, a startup lost all of its data because one Google system expected another Google system to implement a fraud detection measure, and the other Google system didn't. Completely internal to Google, but poof, all of a sudden, account gone, all data gone, and no way to fix it.
In another case, GCE lost a contract worth millions of dollars to AWS because Youtube had algorithm bugs. Poof. Youtube system broke. This was early GCE days, and they were looking for successes, and we were sufficiently high-profile that we had a dedicated engineering team. They pushed on Youtube. Youtube said we weren't important enough to help.
Most of this stuff is either under NDA, or otherwise just private, but it is why my current employer, as a policy, doesn't do business with Google. That's a big contract too; I'm not at a startup right now, but at a big company.
> And name me one other company that provides unlimited free storage
I think you're unintentionally arguing I shouldn't rely on Google-unique features, since I might lose them. I already knew that. I've learned that lesson painfully over, and over, and over, and over, and over.
That hurts Google in B2B. If all customers take that philosophy, Google fundamentally can't have unique, differentiating advantages beyond price. It doesn't matter what unique AI algorithm Googlers come up with for the Google Cloud. I won't use it if I can't count on it still being around when I ship my product. And I can't. The only time I will use Google is if it provides unique customer access (e.g. Android apps or Youtube eyeballs), which hopefully the antitrust thing will help with.
I'm not happy to pay to be the product, or to have something I rely on swept out from under me.
And as for killed services with a billion users, well, free unlimited Google Photos is now a killed service.
As for killed paid services, Google Play Music, Nest Secure, Google Photos Print, Google Audio Ads, and a ton of others.
Google isn't making that case because Youtube is YOUtube. The whole point is that it's a platform for family videos, amateurs, and all sorts of other things.
If the RIAA didn't want their content Youtube-dl'able, they should have used one of their locked-down platforms. They went to Youtube because the locked-down corporate platforms don't have the eyeballs. Now, they're trying to convert Youtube into one of those platforms.
My expectation is the RIAA will prevail, but because of deeper pockets and the potential for a multi-million dollar legal battle, not because they're on solid legal ground.
Perhaps it was inevitable that RMS would be taken down by neutral third parties, but now we'll never have any way to know. The reality is he WAS taken down by people who had been smearing him with exaggerations and half-truths for at least the last quarter-century, and this was part of that dynamic.
I don't feel bad if people are taken down fairly. I only feel a little bad if people are taken down by, in your words, "a witch-hunt by prudish SJWs." I feel very bad when people are taken down by enemies spreading half-truths.
What's curious is:
1) Going against social mores tends to cause psychological harm
2) Looking across history, plenty of cultures broke those mores without psychological harm
That's not just sexual mores. All sorts of things. Corporal punishment is harmful to kids in Western Culture, but not harmful to kids in many other cultures worldwide (yes, with good evidence).
The evidence for what constitutes healthy person-person relationships is mostly culturally-situated, whether that's parent-child, sibling, friends, work environment, or otherwise. The same behavior which might be considered a child abusing their parent in one country might be considered the parent abusing the child in another (and lead to actual psychological harm for the abused party).
It's helpful to look at mores across cultures and history to see which things are fundamental to humans, and which are culturally-situated. Of course, that doesn't make breaking culturally-situated norms okay -- it still causes psychological harm.
What is more problematic is that we can't talk about this in more than abstracts. I could pick dozens of concrete examples, and if I raised them in a public forum, my career would be dead.
My experience with people going after corruption (n>3):
No one is perfect. Everyone breaks some laws, does something taboo, holds some controversial view, or does something improper. If you're going after power structures, they'll take that and pull on it to destroy you unless you back down.
This was done to an extreme in Communist regimes. For the most part, people didn't go to prison for criticizing the government. People who criticized the government went to prison for some other, complete inane but technically correct reason.
For a public examples, See Hunter Biden. Compare to Donald Trump. Yes, there appears to be corruption with Hunter. No, it's not out-of-mainstream for DC. Yes, it's greater than Trump's family. Only Biden has the resources to defend himself -- half the country is supporting him. Unless you're a presidential candidate, something like that WILL take you down.
Allowing that to work really benefits the power structures who can do PR and hire investigators. And it really harms anyone out of the mainstream or with unpopular views, where a simple cancel culture can kill your career.
In some ways, that made him less effective as a free software advocate, as did many of his other more controversial views on just about every topic. In others, that made him more honest.
That's not what's being reported, though.
Median Indians, Africans, Chinese, Arabs, and Eastern Europeans all feel "creepy" to American women (at least until they learn the differences in body language, communication patterns, and so on). Indeed, they feel creepy to each other. For that matter, lawyers, engineers, and marketing people all feel "creepy" to each other until they build those cultural bridges.
"Richard Stallman resigns from FSF, MIT after defending child rape" - OSNews
Seriously. Yes, he's awkward around women, and has sometimes hit on women in sometimes cringeworthy ways. He's also awkward around men, for that reason. There's a difference like night and day between the actual issues, which are about on-par for much of the population, and what's reported. Heck, I'd say that's about on-par for about half of the MIT student body back when it was a genuine nerd school rather than the "#1 branded university in the world."
But he knows that no means no, he doesn't mean harm, and he doesn't abuse his power.
And compare the treatment he gets to other people in position of power. Compare that to the MANY MIT faculty who DO use power dynamics to abuse female students, and who DID fly to Epstein's island to party. Heck, compare that to executives and politicians, or, well, many others in positions of power. You're not going to convince me Stallman is worse than the median there, or even the 25th percentile. It's just that one group is politically savvy, and the other, socially awkward.
And this is exactly how you get a plausible smear. You start with a smaller issue, and you twist and exaggerate it until you have a pedophile-rapist. And yes, it was intentional.
It's unlikely transcoding a video added anything creative to the process.
RMS has lots of enemies, and most of the stuff you read about him is fabricated. You start with a grain of truth, mix in a few half-truths, add a bold-faced lie, and you've got a great smear.
I mean, he doesn't have the best human interface, but people take that, and use that as the basis of rumors to destroy him.
Read about his departure from MIT/FSF. Compare what he wrote to what he was accused of writing.
He's devoted his life to letting people know, and so few listen.
A lot of intermediate approaches -- such as firewalling business units at Google -- would achieve similar effects.
I didn't read this as "do nothing." I read this as "think through consequences."
One of the reason for the decline of the term "dark ages" is that there is increasing evidence that people in this period didn't actually crawl under a rock of ignorance. Large-scale armies and infrastructure projects were gone, but at the same time, a lot more knowledge persisted than is given credit for -- the sorts of things which don't require a massive Roman central government administration to maintain.
Mozilla is (1) 100% dependent on Google for it's survival (2) competing with Google in the browser space, with unfair tactics by Google (3) Critical to the free Internet.
That places them in a tough position.
I suspect if they didn't have unfair competition from Chrome, they could have more marketshare, and ultimately, more revenue from Bing/DDG/etc. (perhaps at lower per-user prices).
Galen was a pretty accomplished medical researcher, and Roman-era medical implements were quite sophisticated, and were well-established by the time of the dark ages. By the late dark ages, we also saw a lot of modern medical techniques coming in from the Middle East.
There's this stereotype of people being filthy, ignorant, and dumb, but it's not really true. And for many of the bad practices from then do persist in poorer parts of the world today affecting likely around a billion people (and some, like homeopathy, even in wealthy parts).
Note that I'm not weighing in on the overall debate about where modern medicine is relative to the dark ages. We've made a lot of progress. I'm just refuting the stereotype of the dark ages.
You don't even need to go that complex. Even moderately complex JOINs start to look better in SQL than in ORMs.
If you have an employees database, an inventory database, and a clients database, an ORM is perfect. If you have a fixed hierarchy, ORM works great too. It maps objects to the database. That's 90% of web apps.
If you're building e.g. an online CAD system with a complex data model for storing hierarchical layers of objects with complex relations, and where you need to perform complex operations on that data for e.g. simulations and optimizations, SQL will actually handle that just fine. You'll just be doing SQL beyond what fits into an ORM. At that point, if you use an ORM, you'll be doing complex contortions. Think back to your data structures class. Then to the grad-level data structures class. Most of those structures, SQL will handle fine, but ORMs won't.
SQL has also turned out to be surprisingly resilient to different programming paradigms. SQL came out just around a half-century ago, before OO was common, and did fine with OO, functional, structured, and a whole range of other paradigms which have moved into and out of vogue over that time. ORMs, as the name implies, are specific to OO.
A few posts up, the poster is correct that there are design patterns around ORMs which work well, and design patterns around SQL which work well. You want to pick one and stick to it. The mess comes in when you mix layers of abstractions. One or two SQL procedures might not kill you, but when you have big chunks of code using ORM and big chunks NOT using ORM, you'll crash-and-burn.
Footnote: "Complex domains" is also about the database layer, and the type of complexity. Right now, I'm working on a very algorithmically complex system, but that complexity isn't in the data layer. Most of the data engineering is about moving GB of data around in realtime. All the database needs to handle are simple things like auth/auth. That's an ideal use-case for an ORM. I've also built systems with ORMs where the database layer was complex, but complex in ways which aligned well with ORMs. So this shouldn't be read as "dumb programmers who can't handle functional use ORMS."
Footnote to posterity: Should someone stumble upon this in a web search: If you grew up in OO and Java, and never ventured beyond, none of the above will make sense to you. You only learned the programming paradigm ORMs were designed for.
And with diets and diseases, if something has a short-horizon effect, that's good evidence it might have a long time-horizon effect too.
The seriousness with which I respond to COVID19 is mostly due to short time-horizon effects giving strong implications of possible long time-horizon effects. I'm less concerned about the 3% CFR as I am about the larger portion of individuals who might have long-term disability, lower economic output, etc. If that passes, we'll be feeling that for many decades.
A few legally naive organizations avoid AGPL due to perceived liability. I have yet to see a case with that million-dollar payout. Perhaps lightening will strike and you'll be the first, or perhaps there's minimal risk because that's not how laws work.
(Disclaimer: Above applies to common law jurisdictions, like GB and US; I have no idea about French legislative law, Asian courts, etc.).
You can write fast native you. You can write fast code by calling into native. The two rarely cross.
If we were clever, we could even have one front-end and multiple backends to go into fast Python, CPython, Cython, embedded micropython, etc. If we were double-plus-clever, we'd add GPGPU, multicore, and compute cluster to the mix.
We'd need someone with deep pockets and high risk tolerance to pull it off. It's a big change. Big changes often fail. Still, when I look at the amount of Python at Google... Well, it'd sure be a pity if all of that became legacy code when we all jumped ship to Julia.
* The Nano's 0.5 TFLOP of CUDA is insufficient; AND
* Google's 4 TOP of DL performance IS sufficient; BUT
* You're not bottlenecked by things like USB2; AND
* You can't afford a higher-end Jetson with dedicated ML
Well, let me just say I can't think of one on things I've actually done. I could come up with ones in abstract, but it seems like Google is aiming for 1% of the market at best. With 1% of the market, NVidia will win on community, tooling, R&D budget for v2, etc., so in the end, that means 0.1% of the market.
And that's if the two businesses were starting on even footing. They're not. Google has a horrible reputation for leaving customers high-and-dry, while NVidia's reputation is pretty good for B2B. Even if the boards were equivalent, most would pick NVidia based on that alone. Plus, NVidia is coming in with established users; they have a first-comer advantaged.
Getting late to the game with a generally inferior product with backing from a less reliable company? It seems like a loser to me.
Seems like a total loser compared to the NVidia Jetson line. It does have a few more TOPs than the Nano, I guess? And it costs a bit less than the Xavier series?
I don't think it needs to be big guy, but it needs to be better than what else is out there. That's a high bar. There are plenty of good tools.
Early Google: Phenomenal ability to build effective, high-quality products. Everyone wants to work there. Stainless reputation with the public.
Current Google: Limited ability to build effective high-quality products (but retty good at sustaining the successful products they have). Mixed reputation as an employer. Mixed reputation with the public.
You can extrapolate from there. I'm concerned about the second derivative.
I think the second piece there is employee growth. In 2001, Google had 300. In 2004, it has 3,000. By 2011, it had 32,000. today, it's over 100,000. You can't quite fit an exponent to it, but it's pretty close. If revenues don't keep growing exponentially....
I'm not implying it's failing, by any means, but it's gone to the same place as any other large corporation, and not a strong position to maintain pole position from.
On the whole, it actually worked very well. Everyone wanted to work there because you were surrounded by top people.
It's not the fastest algorithm (by big-O; probably is in practice), though. You can do a matrix decomposition, exponentiate the eigenvalues, and convert back. Speed will depend on how you do the exponentiation, and that's gets into a pile of optimized numerical methods.