

- encapsulation and meaningful function names.
Oh yes. Deepseek can quote from copyright sources. So can openAI models, but they are programmed not to.
Facebook trained on the torrent of Annas archive.
The copyright horse has left the stable.
No, taxes implies a monopoly on the training data. The government profits. The rights holders get nothing back.
If private data is deemed public for AI training then the results of that training (code+weights+source list) should also be deemed public.
I’ve not found the small models to be good enough to be useful.
I have never and will never interact with my phone by speaking to it and I don’t want to be around other people who are doing that.
Out of context this statement is hilarious.
It used to be that speaking to a phone was the only way to interact with it.
Typing and tapping would soon be passé,
The tech certainly isn’t ready for this. My voice input to chatgpt gets automatically translated into Welsh.
I asked chatGPT about a quote from Iain Banks - The Player of Games. It claims not to know about it’s contents except for the cover blurb. Bullshit.
I fed it a detail and it suddenly remembered.
They must have programmed chatgpt to deny that it has read copyright works.
Deepseek had no such qualms. It couldn’t give an exact quote but it did give what it called an approximation.
What is amazing is how our universe and existence seems to be governed by a few physics laws (which we don’t fully understand).
I think the bug was that a splash of beer had the same alcohol content as a cup.
I see. So dod could be the same across multiple goals.
How is that different from a goal?
Not being argumentative. Just seeing if I’m missing something.
Doesn’t a definition of done count as a goal?