Is this "don't vibe" code, or "don't accept" code written with agents? We have a philosophical (and potentially legal) aversion to code written by models.
I can appreciate not wanting to fill Codeberg up with potential AI slop, though the effort to validate is likely higher than just accepting it. What I think will backfire is being one of the most prominent alternatives to GitHub and prohibiting code written by models, which is currently a large constituency and one which is growing exponentially.
This is becoming a wedge issue. It should not be. As an industry, we can solve this. As an industry, we have too. If we don't, legislators will do it for us. And they'll make a bad job of it. And if you petition your local legislator wherever yiu are in the world, then that's cool, but if this is solved locally, we will see serious fragmentation. As an industry projecting ones politics isn't going to make much difference.
I think the creeping invasion of privacy argument is backwards here. What we have today isn’t privacy, it’s abdication. Platforms are externalising risk onto parents and pretending the internet is exempt from the safeguards we accept everywhere else.
Either the tech industry solves this, or governments will. That’s not ideology, it’s capitalism. If we don’t build workable, privacy-preserving primitives, regulation will arrive in the most blunt form possible.
There’s a reasonable middle ground. Identity can be a first-class citizen without being leaked to every website. I don’t need to hand over my name, address, or documents to prove I’m over 18. I need a yes/no assertion.
Imagine the browser exposing a capability like:
> “This site requires age verification. Are you over 18?”
The browser checks via a trusted third party credential and returns a boolean. No DOB. No tracking. No persistent identifier. Just a capability check, much closer to how physical ID works than today’s data-harvesting mess.
As a parent, I already police my kids as best I can, and it’s imperfect. But the offline world has friction and gates: bars check ID, cinemas enforce ratings, shops refuse sales. Those mitigations don’t make parents redundant; they support them.
Online, we’ve chosen to pretend none of that is possible. That’s not a principled privacy stance.
If we don’t design these primitives ourselves, we will get crude, insecure age databases, mandatory uploads of passports, or blanket bans instead. This is the least bad option, not a slippery slope. Collectively we have solved far harder problems.
I think we need to accept that age verification makes the internet safer. What we cannot accept is age verification's use as a mechanism to pry too far into peoples lives. When we can separate age verification from who am I, most people will be happier. What's tricky is who validates age? Your ISP? Your government? Your OS? A thirty party? Who accredits third-parties, and can you trust them? I'm convinced there's a way to solve this do we can keep the internet safe and not intrude massively on peoples privacy.
That's an epic polemic. If the cost of operating in Italy isn't profitable, exit Italy. If it is, then adhere to the laws of Italy. If Italy makes the cost of business too high they'll dial it back.
But in all seriousness, LLMs have their strengths but we’re all wasting tokens and burning the planet unnecessarily getting LLMs to work so inefficiently. Use the best tool for the job; make the tools easier to use by LLMs. This mantra is applicable generally. Not just for coding.
Microsoft are really sweating GitHub now aren't they? It wouldn't be so bad if it improving but there is certainly a perception that it is costing more for a poorer product, irrespective of the new features they're layering on.
A UK organisation with treasonous, multi-generational experience, that's cited in the article, that people refuse to read or believe? Thanks for re-sharing <3
After 2008, others pressed Keynesian stimulus. The UK chose Hayek. Austerity. Councils took the hit. Services vanished. Early-years centres. Youth work. Local welfare. The safety net thinned, then tore. Families slipped through.
Then Covid. Then Ukraine. Prices surged. Wages didn’t. A decade of inflation stacked up while pay stood still. For many, that was a silent pay cut.
Truss turned strain into crisis. Unfunded tax cuts. Markets panicked. Gilt yields spiked. Mortgage costs jumped overnight. Another blow to households already on the edge.
So we end up where CNN reports: record child poverty, even among full-time workers; parents unable to cover the basics as the social architecture collapses.
Into that anger steps Reform UK. They offer a protest vote. But their plan is the same old mix: deep cuts, a smaller state, and migration as the scapegoat. The very recipe that helped bring us here.
I'm using sqlglot to parse hundreds of old mysql back up files to find diffs of schemas. The joys of legacy code. I've found hypothesis to be super helpful for tightening up my parser. I've identified properties (invariants) and built some strategies. I can now generate many more permutations of DDL than I'd thought of before. And I have absolutely confidence in what I'm running.
I started off TDD covered the basics. Learned what I needed to learn about the files I'm dealing with, edge cases, sqlglot and then I moved onto Hypothesis for extra confidence.
I'm curious to see if it'll help with commands for APIs. I nothing else it'll help me appreciate how liberal my API is when perhaps I don't want it to be?
I agree with this. cat is great for "cating" bat is great for throwing shit on the terminal in a fashion that makes it semantically easier to reason with, two different use cases.
I’m on a Mac, and some of the default tooling feels dated: GNU coreutils and friends are often stuck around mid-2000s versions. Rather than replace or fight against the system tools, I supplement them with a few extras. Honestly, most are marginal upgrades over what macOS ships with, except for fzf, which is a huge productivity boost. Fuzzy-finding through my shell history or using interactive autocompletion makes a noticeable difference day to day.
I can appreciate not wanting to fill Codeberg up with potential AI slop, though the effort to validate is likely higher than just accepting it. What I think will backfire is being one of the most prominent alternatives to GitHub and prohibiting code written by models, which is currently a large constituency and one which is growing exponentially.
Co-Authored-By: Claude Sonnet 4.6 <[email protected]>