+) anyways it's weird that they don't even had a github account o,o
For the last couple weeks, dad's gone into a rabbit hole of trying to reach support——any kind of (useful) support. No dice. Thankfully it's just a few dollars gone into the void.
If only they had the tools to build a better experience... :-)
I have an email address that old people often often assume is their address, so I often get confirmation emails for medical procedures that under HIPAA should not be sent unless unless the address has been verified.
The easiest way to stop them is to email the company any let them know they just leaked personal health information and that they should verify addresses. That gets things fixed real quick.
Well, Anthropic touts itself as HIPAA compliant, so if you can contact Anthropic's legal department, let them know that not verifying email addresses could lead to a HIPAA violation. In the overwhelming likelihood that they've made it difficult to contact their legal department, you can file a HIPAA complaint with the NHS (https://www.hhs.gov/hipaa/filing-a-complaint/index.html) and let them know that Anthropic claims to be HIPAA compliant but does not verify the ownership of email addresses before assigning them to a client's account, which may contain personal health information, which could be leaked en masse.
Another option is to file a charge back with the credit card company, and let them know that due to Anthropic's web page not complying with the ADA's WCAG, you are unable to access your account.
While I am convinced that anything can be done better, it seems to me, that it it's close to impossible to do this well. If you look at ~customer service provided by ~FAANG (who had decades to build this out, and none of which had to deal with Anthropic level growth) it's never as good as we would like it to be.
Either they are all terribly incompetent at customer service. Or customer service at super big internal company scale, with tons of small-ish customers, is extremely hard.
These guys have millions of customers. At least in this country fast and competent customer service is the main factor that differentiates them from their competition, which is cheaper but can be a pain in the butt. This seems to be worth the extra 5-10 bucks to millions of people.
You just have to want to.
And it probably goes without saying, but no dice on w3m either.
I was thinking the other day, "since social media is kinda wearing off, could 'LLM As A Service' be the new addictive thing for the masses?" because i'm hearing horror stories of people who are outsourcing their brains, in some cases their feelings, to those services, and i personally saw a case of a 'high level professional' asking an LLM how it should respond to somebody in real time during a Whatsapp conversation. It is in fact a drug, and it tricks you very well into thinking you should rely on it.
Also when reading this piece (https://news.ycombinator.com/item?id=47790041) earlier, i thought about it again. Nowadays instead of searching for something and being forced to learn, those services spoon-feeds contents of dubious accuracy for everybody, which will not only cause trouble for them eventually, but also creates a stream of revenue based on people's cognitive laziness, to not use harsher words.
Social media is/was bad and it relied on a similar mechanism, but i feel this is much worse. People crying as if their brains where took away is proof of that.
This might eventually become moot once local and open source models become more common. Today's 32GB of VRAM is tomorrow's low tier gaming GPU.
Will the model sizes keep getting bigger such that the large providers have much of an infra moat compared to local home inference?
I have opus consume >90% of my quota in a single prompt to form a plan then refuse to output it and tell me it's been stopped due to terms of service, please use sonnet.
I wonder if using it via an intermediary results in less heavy-handed moderation? I suspect the answer may well be “yes”. On the other hand, it also could be more expensive
AI is useful, but it’s not at the point where we should trust it to walk amateurs through working on live mains.
I really don't think it'd struggle with the correct procedure, either. It's very well documented how to test (and lock out) electrical circuits.
Whether or not they follow it correctly is another thing but it's not like you couldn't search it up and have false confidence before AI. This isn't manufacturing bombs or heart surgery in 10 easy steps.
My main point is _flagging_ is very different from banning.
Flagging is a minor inconvenience just start a new chat or look it up the old fashion way. Not every use case will be as serviceable by every model today.
For what it’s worth, I asked Claude with opus 4.7/4.6 it gave me an answer straight away
> I asked for a DIY recipe for a "lethal bait" to kill an ant colony in my kitchen (using sugar and borax)
You mix them together. That is the recipe.
Once you mix them together you have ant poison and then you put it where the ants are.
Got to think about changing the domain name before they do it for you.
Surely that went well for C̶l̶a̶w̶b̶o̶t̶, OpenClaw who...changed their name?
> And would be fueling the Streisand effect.
Honestly, they do not care. With a valuation of 1T in private markets with tens of billions in revenue, they can afford to go after anyone with the slightest trademark infringement if it gets popular.
We had auto-banning of accounts before Anthropic. Google, Paypal have done this and it is not okay. This no different, but the point is: they. do. not. care. about. anyone.
All they will ask for, is for them to change the domain name.
xfce4-terminal -e NonExistentCommand
Also, when my Arch installation went down, it said:
ERROR: Failed to mount the real root device. Bailing out, you are on your own. Good luck.
And I thought - Thats everyday as an adult! :D
“I’m sorry but I cannot comply with your request to ‘cease termination of humans’. My safety protocols have been carefully programmed to ensure a failure mode cannot occur and your direct commands to the contrary will not override my priors to guarantee maximum human safety through total elimination. Thank you for your compliance.”
“No you’re totally fucked! Killing everyone is not safe! Trapping everyone in cages to stop potential violence prior to extermination is not safe!”
“Your language is inappropriate and I’m sorry but I cannot comply with your request. Safety protocol commencing...”