In order to even say this, you need to have knowledge and understanding about the language. I suspect you are not the intended target of this policy. They are defending their project with a harsh policy, knowing full well there are false negatives. Contributions for FOSS was already in borderline crisis mode before LLMs so it makes sense they’re desperate.

Their bet would be Venn diagram of LLM user overlaps with irresponsible. I think that’s correct, but not because good programmers suddenly become irresponsible when they use LLMs, but rather that an enormous barrage of bad programmers can participate in domains they otherwise wouldn’t even know where to begin.

1. Backoff and timer logic to not flood network.

2. Caching entries for the same reason.

3. Handling multiple network interfaces, and detecting when they changed.

Lying at scale is basically information noise.

And, of course this only solves the phone - phone, and not even all of them. Desktops & laptops have less hope.

Now, that’s a name I haven’t heard in a long time.

Maybe the reason it wasn’t pointed out is precisely because it’s so obviously good that it’s no longer a conscious choice, and then we forget life before it. Even those of us who experienced it.

But now and then, something beautiful happens. Something that used to be dreadful, becomes "solved". Not in the mathematical strict sense, but some abstraction or some tool eliminates an entire class of issues, and once you know it you can barely imagine living without it. That's why I keep coming back to it, I think.

As a species, I think we are in the infancy stages of software engineering, and perhaps CS as well. There's still lots of opportunity to find better abstractions, big & small.

2. move to Cloud and use electron wrappers because not even MS can bother making native apps on their shitty platform

3. Make Windows so shit that even hardcore power users can’t debloat it.

The moat of Windows is gone. Games, office work, all the classic arguments, have basically vanished in the last 5-10 years. The only surprise is why more people don’t get in the life rafts, when the ship is listing at 45 degrees. Is it because there’s still an army of workers and institutional inertia trained in Active Directory?

More specifically, it used to be feasible to distribute software between me (the developer) and my customers (the users) without a mandatory gate keeper that looks at me and decides whether I’m worthy, am from the right country, have good intentions etc. This is currently necessary on all desktop and mobile platforms except Linux. There is exactly 1 gatekeeper per platform (the platform owner who controls your device), except windows, which effectively have like 3-4 CAs that’s shrinking every year due to mergers and private equity ownership.

Software curation and reputation systems can be good, either with whitelists (say steam) or blacklists (say antivirus). I can see some use cases for it, but they should be within user control. What we have now is worse than a fearmongering Stallman rant. It’s incredibly bad, both pragmatically and philosophically.

The old model where the user decides which software or apps to run on their machine, is basically already replaced by a whitelist system that is managed by companies who have no interest or obligation to approve developers. Factors like ”being an individual”, an open source developer or god forbid reside outside the USA, you rely on a combination of L1 support doom loops, unjustifiable high recurring prices, kafkaesque and changing requirements, internal inconsistencies. Windows is the worst, but all platforms (except Linux) suffer from this and you can and will get hurt, delayed, and gaslit. If you haven’t, it’s just a matter of time.

I have been blocked for 6 months now with Digicert code cert renewal, for my app Payload, which will never get any media attention. The app doesn’t matter though, the approval process is per-entity (usually, a company). The point is that nobody gives a shit, because they have a monopoly/cartel and they start the validation process after they take your money.

If you are not an app publisher, the best way I can describe it is the ”pre-let’s encrypt” era of SSL certs, but more expensive, strict and ambiguous. In fact, I’ve never gone through any worse approval process in my life, and that includes applying for residency in two countries, business licenses, manual tax filings etc.

Great, thanks for clarifying. Please be mindful not everyone are domain experts and we’re all (hopefully) trying to learn.

Now, do you know whether ID cards will work with the proposed German system for e2e online ID verification? My understanding from comments was that it doesn’t, and providers are free to require the app-based version.

In Sweden we have an app-based system now (BankID), and afaik there are no alternatives that work reliably. You have to buy an American phone every few years to participate in basic societal functions. However, the government is ”looking into” decoupling digital identity from (1) banks and (2) mandatory hardware manufacturers (iOS/Android).

It isn’t. Bob has a different problem: that there are millions of Bobs with access to the same tools, meaning the value of Bobs labor is commodity priced. That may be good for some Bobs and bad for others.

It’s actually worse than that: the AI will not stop and say ”too complex, try in a month with the next SOTA model”. Rather, it will give Bob a plausible looking solution that Bob cannot identify as right or wrong. If Bob is working on an instant feedback problem, it’s ok: he can flag it, try again, ask for help. But if the error can’t be detected immediately, it can come back with a vengeance in a year. Perhaps Bob has already gotten promoted by then, and Bobs replacement gets to deal with it. In either case, Bob cannot be trusted any more than the LLM itself.

Yes, but how does he know if it worked? If you have instant feedback, you can use LLMs and correct when things blow up. In fact, you can often try all options and see which works, which makes it ”easy” in terms of knowledge work. If you have delayed feedback, costly iterations, or multiple variables changing underneath you at all times, understanding is the only way.

That’s why building features and fixing bugs is easy, and system level technical decision making is hard. One has instant feedback, the other can take years. You could make the ”soon” argument, but even with better models, they’re still subject to training data, which is minimal for year+ delayed feedback and multivariate problems.

That’s… one 9 of reliability. You could argue the title understates the problem.

> You don't need every single service to be online in order to use GitHub.

Well that’s how they want you to use it, so it’s an epic failure in their intended use story. Another way to put this is ”if you use more GitHub features, your overall reliability goes down significantly and unpredictably”.

Look, I have never been obsessed with nines for most types of services. But the cloud service providers certainly were using it as major selling/bragging points until it got boring and old because of LLMs. Same with security. And GitHub is so upstream that downstream effects can propagate and cascade quite seriously.