Here, here! LLMs are 100x smaller than their training sets, they can't possibly memorize them, so what they can do is learn generic compressed patterns. That is precisely what copyright should not protect - abstractions - and instead focus just on expression. LLMs are not even intended to reproduce expression, that is hard and inefficient. It costs money, comes out wrong, and is slow compare to ... copying which is fast, perfect and free. LLMs are precisely fit to be free of copyright issues but somehow people still scream "stealing". No, if you are talking about abstractions, those should be free to "steal".

Defenders of reddit rights don't care about your rights, they just want the AI companies to pay up

Technical solution is to avoid multiple parties doing their separate crawls. If there was a central cache, it could be set up such that crawlers pay the cache, and the cache pays the websites, ensuring fast and fresh access.

The crawler is representing a real human this time, it should get the same rights as the real human. Consider it an assistive technology, like TTS or Braille reader. The best way would be to just use a Chrome tab to perform the operations using a Computer Use agent.

Your system would collapse internet usage under connection costs. Not just bad bots, but also good usage would decline.

> How did we all decide to use the phrase vibe code?

Andrej Karpathy came up with it : https://ancillary-proxy.atarimworker.io?url=https%3A%2F%2Fx.com%2Fkarpathy%2Fstatus%2F...

I bet every time you used Google you "mined" a dataset containing "every bit of data". So get off your soapbox.

Such a layoff of skilled programmers that can successfully automate with AI means they get hired immediately.

NYT is making copyright a weapon against creative freedom. With each new work they publish, they reduce our freedom to create. Copyrights last for too long, and have been changed from protecting expression to protecting abstractions. The AFC (abstraction filtration comparison) test and the "total concept and feel" interpretations basically make it impossible to do due diligence, nobody can ensure they don't infringe any abstract idea in all the published works. The law is biased against new creatives and public domain. In fact copyright is extremely outdated, we prefer cooperation, sharing, and doing things together now, and that doesn't work when everyone holds to their royalties like NYT. And focusing on AI infringement is far fetched, AI is by design not a replication machine, copying is a million times easier to achieve already. AI is used for novel outputs.

They were ordered to keep logs of API calls which were not subject to logging prior to this decision. The web interface has different logging.

If you punch someone in the noise - you get part of the copyrights. It's already both a work of nature and man!

Let BBC hold copyright over their expression while Perplexity serves an analysis from multiple sources, which is not identical to any one of them and provides an transformative output.

The market for people who want to read random 42% snippets of Harry Potter will collapse.

LLaMA might not have the mitigations, but OpenAI, Anthropic and Google filter the outputs to ensure no regurgitation. It's not a problem on the output unless we count abstractions as infringement.

It's been only a short while, but good by Slack MCP.