what a bad idea...

this would be laughably easy to circumvent and terribly easy to abuse

study says we should do a real study

Yes, because you need evidence from small studies to justify running larger ones. You're not going to run an n=1000 experiment from a mere guess.

I think the frustration is: it's not "news"

Help me understand the line Anthropic is drawing in the sand?

Don't get me wrong i'm glad they are unwilling to do certain things...

but to me it also seems a little ironic that Anthropic literally is partnered with Palantir which already mass surveills the US. Claude was used in the operation in Venezuala.

Their line not to cross seems absurdly thin?

Or there is something mega scary thats already much worse they were asked to do which we dont know about I guess.

The whole reason this is happening is because Anthropic looked into how Claude was used in the Maduro op and found it to violate the negotiated terms of service.

Their hard lines are:

- no usage of AI to commit murder WITHOUT a human in the loop

- no usage of AI for domestic mass surveillance

So... this would be fine with them?

Claude: "Are you sure you want me to commit murder?"

User: "Yes"

Or do you mean Human presses button:

Claude: "Do you to commit murder? If so press the button."

User: "I pressed the button"

Claude: "Great! Now lets summarize what we did."

First one

Seems like an absurd distinction to me... Reminds me of "I was just following orders"...

I mean the distinction doesn't really matter

There are many ways to construct HITL UXes. But typically they'd take the form of the first one

I think you're missing the forest for the trees. All Anthropic is saying is that HITL is required before murder, the UX is irrelevant

I agree the distinction doesn't matter, but im not so sure "just" having a human in the loop qualifies as an ethical stand. Just because your not pulling the trigger doesn't make you not culpible for the outcome.

I don't understand the line as well. So its no to domestic surveillance, but all other countries are a fair game? How is this an ethical stand? What sort of mental gymnastics allow Anthropic to classify this as an ethical stance?

To me all of this reads like "we don't trust our models enough yet to not cause domestic havoc, all other is fine, and we don't trust our models enough yet to not vibe-kill people". Key word being "yet".

"vibe-kill" made me laugh then feel sick

Always open to an opinion if its backed up by deep analysis and some well presented facts but this whole blog is just short angry comments.

I don’t think you need deep analysis to be skeptical that this next big thing won’t ultimately be enshittified like every other big thing that was once free, powered by VC cash.

What makes this different for me is that I can’t run Uber on my M4 Mac Mini, but there are already plenty of impressive local LLMs that I can run, and the capabilities continue to trend in the right direction

I dont neccesarily disagree I was just dissapointed at the low effort post.

As someone who has been excited at the proposition that we could allow non technical folks to make complex apps without needing to raise money I was interested in reading more than what was presented here.

I would be fine with this if they fixed the wildly buggy browser extensions.

The number of times my preferences have been wiped after an extension update is maddening.

Considering you can very easily be forced to login with your fingerprint by cops, bad actors etc... I would (personally) avoid doing this.

Accessing your laptop is one thing, accessing all of your passwords for every app with a fingerprint is scary.

This is very dissapointing as I switched to porkbun recently to avoid using anything the monopoly Crazy Domains (Dreamscape) owns.

"or anonymize it" means they would never opt to delete it and never will. Theres zero financial benefit to them deleting it other than honoring the customers privacy which they have clearly calculated is not profitable for them. Here's hoping this trend massively backfires across industries.

pledging allegence aloud to anything is bizarre and reminds of some sort of knights of the round table cosplay. It's especially weird making kids do it.

Cults doing cult stuff, bizarre indeed and seems widespread, strangely enough.

Completely unsuprising as Clearview has been turning Orwell over in his grave for years.

Facinating that they landed on infinite scrolling as the problem to spend time and energy on instead of all the other things happening online that have an impact on society.

Genuinely curious about the actual data on this.

Does anyone have a link to a reputable, sizable study?