munksbeer

Claude writes probably 95% of our code now, fintech, amongst top 5 in the world in what we do. I am 100% certain we're not even at the forefront of using agents for coding compared to some others.

It definitely can scale.

> We need laws around this stuff.

The EU created laws around this (GDPR). Americans seem to hate them.

I keep seeing this sentiment. Claude Code cli works very well for me.

To lay my cards out, I think the negativity towards it isn't driven by anything objective, but by anti-AI / anti-big-tech sentiment.

In the UK, you have to do KYC checks, and any crypto you buy will be visible to HMRC (taxman). If you dispose of it for monero, this is a taxable event, and they can tell your crypto has moved. You may not get pinged immediately, but HMRC can and do come after people retrospectively.

> Just that. Like insurance, it was not there to make anyone's lives better

This is revisionism. That is not at all how it started.

What did you mean? I can't really understand your first comment, and this one doesn't explain it any further.

It's probably a typo. He meant sealed interface. I think that should be quite obvious.

It sounds like you are speaking to someone who has actually experienced this and you have not. Trying to be civil here, I'll phrase it like this:

Try not to be so judgemental over a keyboard when you have no experience to speak from. Life is complicated.

>Consuming copyrighted material personally is fully legal, training a model on that same copy-righted material is illegal. Where is the contention?

The contention is obvious in your statement. Is it illegal to train a model, and why? Why is it legal to train a human on copyright material and then sell those skills, vs training a computer and selling the skills?

This isn't black and white.

I remember thinking like this about software development several years ago.

"AI will never be able to reason and write code as well as a human".

I was terribly wrong. I assume you will end up being wrong about engineering too.

I don't think it is easy, otherwise this wouldn't be such a contentious and frequently discussed issue.

A human who trains their brain on material they don't own then creates art or writes code based on this training, sells the product too.

When discussing this, may I ask (I know you are probably bored of the actual arguments), what does "trained models on data that wasn't theirs" actually mean in practice?

Again, I know these arguments have been done to death, but every human who reads source code that wasn't written by them, or views art that wasn't created by them, and practices against this art, is training their brain on data "that wasn't theirs".

They are frequently making a living doing so.

Is this distinction the scale, or is there actually a different more strict definition that we should be using as a common language to talk about this? As in, I should not even be reading certain source code if it is not licensed appropriate, or I will be in breach because I'm training myself illegally? And the same question for art, etc?

I've just realised how infrequently I visit regular websites these days. Everything is trashed, popups, cookie requests, ads splattered everywhere.

I read some blogs that get linked from hn. I read some reddit subs. And I use LLMs for "search" and questions.

Whenever I do have to go back to the regular web it is horrible.

> Cursor is training data and hostile distillation.

What is "hostile distillation" please?

The US has started doing it too.

What happens when teens lives start getting properly ruined because they end up on the dark web or other sites? What about when the first teen commits suicide because they were lured into a honeypot vpn and their data used to blackmail?

Will you still be claiming that this is "good enough"?

>I want it to not be normal to spend all day on TikTok

I would like that too, but I genuinely believe these measures are going to make things worse for children, for the reason I explained.

That includes hackernews right?

You are really so naive to think that banning the big public social media sites is going to stop children from seeing bad things on the internet?

I promise you, within a few years there is going to be an underground social media network or sites that are harder to monitor, don't give a shit what the UK government thinks, and definitely don't care about children's safety at all.

They probably already exist.

AI currently surpasses humans at many tasks, and the list grows every month. My guess is that within 20 years (as a hedging my bets 95% confidence level) AI will surpass our intelligence for almost everything.

The arguments over whether it truly "thinks" or "reasons" will be superfluous. They simply won't matter.

Just absurd. But I think the majority of parent voters love this stuff.

> "iroh is a modular networking stack written in Rust.

You can omit the "written in Rust" part, but then you'd lose street cred.

He is a vested interest, so we can pretty much dismiss what he says.

Even in that post he makes some logical fallacies.

I doubt this will turn out to be true. That seems trivially easy to solve for with correct training.

> The fatalist philosophers and authors have been mostly proven right as time marched on.

In what sense? Most people today would not choose to go an live at any other point in history, except for window between about 1960 and 2008.

Life was much harder for the average person outside of this window at any other point.

> They only emulate reasoning.

So do we.

>Speaking practically your hypothetical is a scenario that requires somebody that is proactively interested in, and theoretically capable of, making a e.g. dangerous virus, yet are unwilling/unable to do so without a chatbot. How many people might this possibly apply to? I think the number is literally zero.

I don't disagree with the rest of your post, but this doesn't seem correct.

I think I'd phrase it that there probably already exist, or will exist, people with the inclination to cause global mass death, but don't have the knowledge or ability to manufacture a virus to achieve this.

Does your team then manually decide the results by going over the PRs? I suppose you know what you're looking for now, but isn't this still quite painful?

Don't feed the trolls Simon.

Ok. I get what you're saying. I definitely agree with losing touch with the code, but I still review everything the agent writes, and I steer it heavily. My perspective is from what I observe. The entire industry (my industry) is embracing agentic coding. And I don't believe developers are doing it because they're stupid.

I think it is going to continue to get better, and I don't think we'll be having this argument in two years time. Our entire industry will look very different.