seek3r00

“I could not stop thinking about the applications of such a technology and how it could improve our lives.

I was thinking of how cool it would be to build a Twitter-like service where the only posts are GPT-3 outputs.”

This could have been either the output of GPT-3 or someone who doesn’t know what they’re saying.

The title is obviously clickbait-y, but it’s fine: they’re trying to sell a product (Google Colab).

IMO if you’re interested in AI research or ML engineering, you already know that — in order to avoid getting people killed - you have to understand how it works under-the-hood. You’re doing yourself, your employer and your fellow humans a favour.

Just keep up the good work, and ignore the bullshit. If an AI winter comes, you’ll be well prepared to migrate to another engineering role.

It reminds me of Coders at Work: https://en.m.wikipedia.org/wiki/Coders_at_work

“To put it simply: the overestimation of technology is closely connected with the underestimation of humans.”

Exactly, that’s what I thought. Although, they should be able to look at the set of instructions that generated the output, which is basically an algorithm in itself. Then they could try to prove whether that algorithm would really generalise.

The model is given the input and a set of instructions (e.g. swap) for producing the output. Essentially, it’s taking a permutation of the instructions that minimise the running time, based on some underlying patterns in the data.

Some interesting extracts:

“The instruction set together with the input representations jointly determine the class of algorithms that are learnable by the neural controller, and we see this as a fruitful avenue for future research akin to how current instruction sets shaped microprocessors.”

“The generalization or correctness we aim for is mostly about generalization to instances of arbitrary sizes.”

“[...] computing a = f(s) can be more expensive on current CPUs than executing typical computation employed in the algorithms studied here. We thus hope that this research will motivate future CPUs to have “Neural Logic Units” to implement such functions f fast and efficiently, effectively extending their instruction set, and making such approaches feasible.”

It makes sense to me. You can use WSL, and run PHP on Linux.

tl;dr: Training learners is becoming cheaper every year, thanks to big tech companies pushing hardware and software.

Well, you’re right, but the hypervisor under-the-hood is better.

Yeah, the thing is that most of us don’t have the patience of figuring out how Nix works , write derivations and stuff.

Docker and Compose look pretty straightforward in comparison, at least for local development.

Look, I get it, Docker has its downsides and it wasn’t meant for local development.

But it just works (most of the times). I can worry about getting shit done, instead of messing around with web servers, interpreters and package managers (it’s fun the first couple of times, but then I want to get to the code).

The problem is that Docker on Mac sucks, because it’s basically a VM running Linux, and it eats up most of my 8 gigs of RAM (yeah I know, 2020).

Maybe, what Apple really needs is something like WSL2.

I meant that we’ve already got Parallels Desktop. I didn’t understand whether they want to introduce something like Hyper-V/WSL on macOS.

They talked about virtualization but they only showed Parallels Desktop. Is just a virtualization layer for x86 on ARM or are they going to provide something like Hyper V on Windows?

Preface

I think that you should always talk about average users. There is no point in talking about technical users that make up a small slice of the market pie.

I’m an Apple customer. I’ve an iPhone but I’m still horrified from needing to buy a developer account (which is like 99$) to install (my) apps on my iPhone (only). It would be probably possible for Apple to enable checks for personal apps, executing apps that are signed with the same iCloud account or something like that (doing all the checks on the device, no servers involved).

Never mind that macOS is a hot mess for (some) developers, right now. So, yeah, I have some issues with Apple too.

On compensation

I see a certain similarity between Hey, Spotify, Restocks. They all are (paid) apps trying to take advantage of Apple user base, without compensating Apple for it.

Meanwhile, Apple is expected to keep paying for the App Store servers, its curation team and its app approval team.

Obviously smartphones are getting better hence (average) users will buy a new phone less often and Apple is expected to keep supporting old devices for 5/6 years and it has to pay the engineers who work on the OSes and the APIs, and the non-engineers that keep the show running.

You could argue that Android and Google are better.

But Android is not Google’s main business, ads are. Google has built (mostly free) services that generate data about its users, which is then used a bait for advertisers.

No one cares if you’re using F-droid instead of the Play Store, Google is still getting its checks from you and your org using Gmail, Analytics, AdSense, YouTube, etc...

On side loading

Side loading apps (for the average user) is risky and it’s mostly done for three reasons IMO:

- cracking apps - accessing apps that are not allowed on the Play Store - escape Google

I don’t really have much to say about side loading. I’m too biased against it, so maybe someone else could convince me that it’s a good thing.

Conclusion

What would happen if Apple allowed apps to use external payment options (like Stripe)?

Well, the obvious, no one would use IAP because there would be much cheaper services that are not compensating Apple.

Apple is far from perfect:

- the App Store approval process is a mess - the communication process is broken (removing old message exchanges, come on...) - they really don’t seem to care about indie developers

However, although Apple needs developers to keep running the show, without no Apple there is no iOS, no App Store. It’s a tricky problem to solve.

The vulnerability should have been disclosed to Tails developers as soon as Hernandez was arrested.

Maybe, AndroidPolice should change its name. Just saying...

I agree with your notes about the connotations of those words. Just keep in mind YMMV depending on which country you live in.

Indeed. I would argue that instead of changing their OSS projects names, people could donate and protest.

And still we let GitHub block Iranian citizens from accessing their repos.

I’m pro BLM, and I’m really outraged by the behaviour of police across the US.

But this has nothing to do with software names, and it is so a US-centric point of view.

That’s true. If this is the case, is it enough to do a good job in growing the business, in order to maintain control of it, or you can do something else to ensure that investors won’t pull the carpet from under your feet?

The problem I see with raising money is that, practically, you’re giving away your business a piece at a time.

Sure, by raising money you’re going to grow your business more easily, but you’ve also compromised your ownership of it.

Obviously, if all you think about is cashing out after the company’s IPO, raising money makes sense. Especially, if you don’t have a sustainable business in the first place, and you’re losing money.

I agree, as long as you can find time to research and understand if you’re really correct. This way you avoid conflicts but you still learn if you were wrong.

No, it would not be my job to decide which discussions should a user be driven to.

If a user is driven to political discussions, so be it.

Sure, this is good for the company because it means the user will spend more time on the platform, but it is a side effect really.

AFAIK, TempleOS seems to be the only operating system running everything in supervisor mode.

I wonder what would happen if we can find a way to build software protection in supervisor mode, ditching user mode and associated context switches.

Ignoring bad actors: what are useful applications of this stuff? It’s not a provocative question, I’m asking for real

Indeed, on the book he says that engineers tend to underrate the importance of the “marketing people”, and that a good product alone won’t sell itself.

In his own words: Thiel prefers to go from zero to one, creating products on which you can build a monopoly.

I think that he agrees on Andreessen about the importance of timing, but he doesn’t consider timing alone a recipe for a successful business.

Also Thiel seems more focused on technology as a whole, rather than the software-centric position adopted by Andreessen.

I think that both views have pros and cons, and that we would benefit from trying to find the greatest common denominator between the two approaches—instead of discarding that or this view.

PS: I read and enjoyed Zero to One, and I’d like to learn more about Andreessen views. If anyone has any suggestion...

I can’t get pass the poor variables naming, though.

Identity verification + Taking in consideration the tickets sent by users being harassed ?