ipython

We had processor level branch predictors. Now do we not only pre fill the next prompt, why not just start generating the response as well?

Interesting thought at least.

About 10 years ago I was browsing Home Depot and found an electronic Bluetooth enabled egg carton. It paired with a mobile app to let you know you were out of eggs. Don’t know what the target market for that product was either.

So the DoW should have the ability to perform mass surveillance on American citizens?

Mind you, that is not the DoW’s charter. How would a restriction on that be “unfit” as you say?

The DoW is perfectly within their rights not to purchase anthropic products. That’s different than labeling them a supply chain risk.

the default is weak because it's the most user-friendly option. Otherwise, you end up with edge cases where the user lost their only device, or they forgot their password, or ... the list goes on and on, and they lose all their previous chat data.

That's a very user-unfriendly place to be. Most users won't understand why their data is gone, become unhappy, and distrust their devices to safely store their data. It's not even about buying more Apple products at that point - it's just literally reinforcing the stigma that "I can't understand tech, it's too complex".

You can't be security-maxxing and user-experience-maxxing at the same time. I do like Apple's approach which at least gives you the option.

I would argue the democrats are more a process oriented party. They wouldn’t pull this shit because they care about process - so much so that they wouldn’t even consider half the stuff that’s going on now as it circumvents process.

In that case, would the people caught unemployed and with upside down loans in the 2008 financial crisis would have a reasonable argument for "force majeure" to cancel their mortgages? ... After all, they had a reasonable expectation that the economy would continue to support their employment, either with their current employer or by finding a new job that paid just as well?

at some point the phrases loses its meaning if it becomes applicable to pretty much anything.

doesn't the reasoning trace provide "intent"?

aaronsw was just a few decades ahead of his time. He should have just been employed by OpenAI and asked a swarm of agents to "download all scientific papers". Because as we have found out agentic systems (and their owners) have zero accountability, unlike humans. Sounds like agents already have more rights than we do.

rip.

All I could think of while reading this was “has no one heard of ntptrace?”

Your comment sparked an idea - it’s like the bike shedding of written text. If you write clean, concise English, it’s easy to parse and then inject your thoughts. If you instead (like an llm) throw a human a large wall of jargon heavy text, they’ll just give up and agree with you.

I mean, I get your thought process and don't disagree. That said...

Would it be an affirmative defense if we had a defendant who said "but your honor, I was told that when I hacked this system, I was operating in a sandbox. I had no idea that I actually had Internet access!"

The frontier is spiky and all, but you have to suspend disbelief quite a bit to, on one hand, have a model that can produce a novel math theory, and on the other hand, that same model can't tell the difference between a "sandbox" and the open Internet.

So, yes, the misalignment had a lot to do with "instructions unclear", but also a lot to do with the fact that the models themselves were not aligned to validate the assumptions and have a healthly level of skepticism, as a real human actor would.

As opposed to the frontier model company that, after discovering that their highly persistent model under test just breached the only thing between it and the open Internet, shrugged and said- let's restart it and keep going!

If anyone looks like the adult in the room after that incident, it's Hugging Face.

Agreed. And they have been poking fun at the right for decades. Heck they made a whole movie about how ridiculous the global war on terror was, back in 2004.