I get tons of ads for these things, or days annotation.

That “job” is just data annotation and almost definitely does not pay what they say.

I wish you the best with your job search. It sucks out there.

Me too

I don’t think the US government has any sort of grand plan more coherent than one of Trumps oily McDonald’s shits anymore.

Just release the files, nobody in the US is going to do a god damn thing anyways.

Yeah a mix of regex and heuristics to validate before parsing

It was a long time ago now

It also had to parse ipv4 because they can be embedded (IIRC) and the different octet formats

How do I short the entire US economy?

Normally I’d say this is the stupidest fucking thing I’ve heard, but I don’t even know if it makes the top 10 this week.

There’s a very high chance of racial bias issues here

Actually, now that I think about it, you don’t even need to tell the model about the ad, you just weight some of the input tokens with the ad embedding and it naturally biases towards talking about that brand.

It won’t be long until the LLMs are trained with for ad-RAG

While each conversation happens a second model will scan it, send the contents to a vector search, return as embedding and a score, then give those outputs silently to the conversing model.

You do need to train/fine tune the model to not mention that it received the ad and subtly push it.

I wrote and ipv6 parser once.

Never again.

Anthropic never promised unlimited though, the plans are Free/Pro/Max/5xMax and it’s branded as “More Usage”

I think they sold too many subscriptions for their capacity.

I bought a monthly sub for this reason. I’m getting 5x my subscription vs paying for the API.

I plan to move to self hosting open models soon, I’m just waiting on a Framework desktop to arrive.

The new Qwen 3 models are decent now and open models are only getting better.

There’s no moat and I get most value from MCP servers, so everything of value is portable.

I’m seeing a lot more discussions moved to GitHub issues and closed off discords than stack overflow :(

Company with incentive to gouge you invents system that gouges people and claims no recourse.

Have people seen the shit that Redditors were doing with Claude?

They literally had a leaderboard of who could spend (waste) the most tokens. There’s a guy running a virtual “simulation” of Venice with according to him 100 simultaneous Claude connections (he has a post claiming they achieved sentience).

It’s also shitty that instead of going after the bad accounts they’re punishing everyone for it — the 5% of users being expected to hit the limits feels way too high.

I know this is the fuck AI community, but I find Claude pretty damn useful, and it’s frustrating as hell watching people abuse and waste something then complain about it.

So far Anthropic have been less shitty than most, though they do have their issues.

I don’t care.

I avoid the services as a protest and I also enjoy the alternatives.

I’m not going to bankrupt them, but I’m not helping them.

The out of the box containerization is still pretty new though — it’s like a month old

At least they’re honest, they did spill tea.

A whole lot of tea.

I’m sure they’ll add some tiny text “this is what your meal might look like” or “not an actual photo”

Just gross.