553 karma · joined February 2, 2022
Adding a shortcut to your homepage is really easy if you're that addicted to HN: https://lifehacker.com/how-to-add-a-website-shortcut-to-your...
Only accept professional emails(i.e name@company.com) from a domain that is older than X days, you can find APIs to find out if a mail is from free providers. + maybe a phone number from the same country of their IPs.
"Sign in with google" is pretty useless, you can bypass that for pennies.
If you're going to require a credit card, you can probably filter them to block debit/prepaid card, or try to make a temporary charge for a few dollars before reverting it.
You can use websites like https://seon.io/ to find out if a mail/phone number looks legit, they check things like if they actually used it somewhere else etc...
How much GPUs time are you going to offer for free? The question you really want to answer, is how do I make registering a fake account more expensive that whatever someone would get from mining crypto with your GPUs.
You can and should of course should block proxys/vpn ips.
Some basic bot protection would not hurt, but it's not an actual solution.
If you don't expect a huge amount of free users at first, you could maybe check registrations manually.
Maybe ask them to sign in with linkedin and check their account legitimacy(number of relations etc...)
An actually viable solution would probably be a mix of multiple of these things, your goal is to make it expensive to register more than one account, not impossible.
Training models on higher resolution would become exponentially more expensive, while still not being that likely to succeed.
And even if you had the schematics, a foundry would still cost millions to setup, + you wouldn't even know if they were any good before trying them out, it would probably be an unpredictable black box.
Codellama is likely a lot less powerful than GPT.
It also happens when I ask him to write class or functions, he puts a lot of ellipsis/placeholders
You just have to make it sound like you could maybe potentially spend money on them one day(instead of just being a curious nerd trying things out)
But when you think about it this way, your brain is also just autocomplete.
You give it input, tons of input, and your next thought/action/sentence is really just what your brain is autocompleting from the previous data he got, your brain could essentially be summarized as a few dozens multimodal gpt-6 running at the same time, interacting with each other, connected to sensors and a few "organic servomotors".
A LLM can be trained to think, and it will essentially autocomplete a thought process, outputting it(thinking out loud), before autocompleting an answer.
ChatGPT was overly trained towards safety because A. a lot of field experts are terrified of what could happen if LLMs were unhinged, B. because they don't want to be sued. C. OpenAI would rather have ChatGPT output dumb answers than dangerous ones in the case the person reading them is naive, overly confident or mentally challenged(or a child).
I like to think of ChatGPT as a naive 6yo telltale nerd with all the world's knowledge who speaks mostly very formally and with the emotional maturity of a toddler.
f** google
So the response is 0$.
You could also write a simple function that turn HTML into a json that only contained innerText, src and hrefs, and ask the LLM to only keep the relevant data
It's rootkits/RATs, just malware developed by intelligence services and/or some technical branches of police, although they sometime hire external contractors for this.
They use exploits or physical access.
AFAIK there is no manufacturer giving backdoors to the french government(but the US and China definitely have some, I wouldn't be surprised if the US shared some access for major cases)
Did you ever have a single history lesson in your life?
Where do you think Afro-americans came from?
Why do you think half the continent is speaking french or English?
Why do you think 14 African countries use a money created and still controlled by France?
Plenty of mines and oil depots are still profiting European countries.
Multiple generations of slave descendants had to pay reparations to France.
Leopold II was chopping hands off Congolese people for rubber.
Using any kind of community coordinated IP ban is useless and would hurt a lot of people, millions(or even billions) of devices have dynamic IP addresses.
You would not stop botnets from DDoSing you and on top of that you'd block millions of legitimate users.
Each of us could unknowingly interact with multiple LLMs everyday which would only have one purpose: manipulate us with a never-seen before success rate at a lower cost than ever.
At some point AI generated content could become more common than human content, while still being indistinguishable.
Good enough automated online propaganda could routinely start (civil)wars or genocides, Facebook already let that happen in the past, manipulating elections would become systematical even in the most democratic countries.
What already happened in those areas in the last few years, is really nothing compared to what could happen without enough regulation or barriers to entry in the next few years.
What's worse is that all of this, would not just be possible, but available to every sociopath on earth, not just the rich ones.
AI regulation is as needed as radioactive material regulation.
Nuclear energy is great, Hiroshima not so much.
You don't need AGI to do damage, current LLMs are already dangerous. IMO, an open-source affordable unfiltered GPT-5 would ruin the internet in a few months.
A ton of legit researchers/experts are scared shitless.
Just spend 5 minutes on EleutherAI discord(which is mostly volunteers, academics, and hobbyists, not lobbyists), read a tiny bit on alignment and you'll be scared too.
There are dozens of worst places around SF, LA, NYC, Chicago. There are no crackheads in Marseille. 100x less guns than in any American city.
But they can't share actual data if everything is encrypted.
Marseille's worst neighborhoods are really chill compared to the bad parts of any American city.
The streets are pretty safe.
Your comment is kinda ridiculous, stop watching TV.
Yeah there are lot of taxes, but those taxes come in handy when you can get 2 years of unemployment benefits when starting a company. Or when you know, get a cancer that would cost a few millions to treat.
even through the API you can't easily use the regular models for chat, the parsing would be atrocious and there are hundreds of edge cases to handle.
ChatGPT4 through the API is the SOTA