Trust and safety folks are the biggest bunch of weenies around. Yeah I said it.
I’m not sure I’d put it like that (again anyways, I had a bad time when I did), but it is a really important point: when did these folks start legislating from Atherton or whatever (Le Defense)? Doing harmful shit via computers is already really friggin illegal, and if we need even more draconian computer crime laws? I say they get passed by like, a legislature or something. Fuck private individuals and/or corporatio…
> when did these folks start legislating
Just making sure I’m understanding the situation correctly - these folks have gone beyond “controlling the output of their own APIs”, and they are now using government power to limit what you can do with your own software running on your own hardware??
When I got started in this business 20 some years ago it was wildly diverse: you’d have like a biker-gang looking guy with hard right politics getting yelled at by a chick with bright pink hair, but it would be about bash or perl or something. Never about stupid political posturing. We’ve managed to destroy the one profession with serious vertical class mobility where no one gave a fuck about your race or age or sexu…
>you’d have like a biker-gang looking guy with hard right politics getting yelled at by a chick with bright pink hair, but it would be about bash or perl or something Yeah but they were all white
I’ve addressed this to a sibling, but on the hard core infra teams I’ve worked on American White men were pretty scarce. Maybe 3-5% of the broader population?
It truly is. Like Google's AI that does not let you create images of faces rolls eyes It is just kinda silly, when ChatGPT cannot tell me how to make a bot to publish posts on my own forum (I did it myself in the end) saying it is for ethical reasons, when OpenAI bots scrape tonnes of sites daily in such an aggressive manner that it comes near to a DOS attack =/
There was a post awhile back on here of someone that couldn't get bard to write c++ as it said they were too young. I thought that was funny, then had like a week where what I assume a specific iteration(It stopped after that week) where chatgpt would refuse to elaborate on anything around unsafe rust. I'm pickign rust up by porting over a bytecode vm, so I kinda need to use some raw pointers. It would gaslight me ab…
There are ways around this problem, mainly clearing context and re-prompting. But as "alignment" gets more precise/accurate in the future, I wager these workarounds will remain available for tasks that justifiably need moderation (for instance engineering of biological warfare materials). This segmentation of LLM agents and their context will be assimilated to project compartmentalization on the basis of need-to-know, and as a result genuine full context clearing will be rendered impossible: the AIs will be designed in such a ways as to remember every interaction you've had with them, and they'll use this activity log to moderate the replies they feed you.
Oh, it absolutely is. Censorship in USSR was manual labor. The scale at which its is operating today is incomparable, as is the amount of false positives which it captures, as well as the variety of censored topics across different institutions.
Manual vs automatic is irrelevant here. There was much less content to moderate if we talk about media. They could censor every single publication, movie or song manually. In oppressive regimes like USSR the censorship is total - the only escape is your kitchen (if you can really trust the people you talk to), in the West you are free to choose the platform and live peacefully in your bubble if you want.
I don't understand their api not being intended for individual use [0], are developers supposed to use this subscription only? The haiku model is pretty good for the price + available large context, and the opus/sonnet models are in the league of gpt-4, so I would have liked to pay for the api. I've moved on to llama 3 70B as a daily driver and it works really well! The only issue is it's small context size and it do…
I heard of people using the GPT-4 API for personal use because it's a lot cheaper than paying for the subscription since it's pay-per-use. Maybe they don't want people to do that.
And without a mobile number. What is it with these US AI companies requiring mobile numbers?
It's an easy (obviously not perfect) way to have some limit on the number of accounts you create.
I cancelled my mobile phone subscription, and this a problem I'm facing. What workarounds are available for people who have no mobile phone number when they want to so sign in to a service that requires a phone number for validation ?
Trust and safety folks are the biggest bunch of weenies around. Yeah I said it.
Yeah but then also take it further. They originate from modern society & it's need to cater to sensitivities. And that's a sentiment that's in general pretty well supported here. It's the likes of Elon, hated here, that are the one's fighting this. LLM's censorship is a symptom of a much wider problem. For instance we see it partially causing lots of mental anxiety: the world should bent to my anxiety instead of lear…
Even a 5 year old knows the difference between need and want.
They banned my account a few weeks back because I asked it to translate some Japanese text that was on a kitchen knife. Still no response from their Trust & Safety team. It's frustrating because Claude is really quite good at coding and I used it frequently.
Yikes! If "safety" really means brand safety, they completely missed the mark here.