4,357 karma · joined September 22, 2013
Invoking the EMH in this context is bizarre, though: if it was fully true, an active fund manager would be as useful as an active roulette manager.
or even a lawyer for that matter
Fix that and it'll provide answers for the rest of your concerns.If you're doing something for those pilot customers that they could sue you about, you want to understand & control your exposure to liability BEFORE something's gone wrong.
https://medium.com/@esperonleticia/transactional-push-the-si...
while people try more and more to regain control
It's not complicated: you turn it off.If you suggest making it impossible to turn off the AI, society turns YOU off.
How in the world can this number of dependencies be considered
sane and normal, and most of all, necessary?
Wait until you find out how many parts are inside an internal combustion engine. I described it: on Tik Tok, on YouTube, on dev.to, here in
another posts. But somehow people don’t say anything.
If you already feel like you've been repeating yourself too much and you don't even have a landing page yet, you may not have the right mindset for starting a company from scratch. Every time you bring up the idea, you need to explain it.I dug up this post from _today_:
https://news.ycombinator.com/item?id=38973944
Some fast reactions:
* this sounds like client certificates but with an extra step involving email
* I don't understand the threat model where you don't trust the user's SSL connection but _do_ trust their email account
they are poised to be increasingly adopted by corporations
Hopium addiction is a terrible thing to see. TELL YOUR CHILDREN! hope the content you want is there, and not AI generated
lolwutThe solution to "content MIGHT be AI-generated" is to switch to content that's explicitly AI-generated?
There is nothing but more data and bigger models standing
in the way of CoPilot getting an order of magnitude better.
So all we need is 100x more code than Github, that should be easy. /s"AI won't make new bioweapons because we already have plenty" is a comically-bad argument.
The anti-LLM panic is going to lead to corporations like Disney owning not just the likenesses of the characters in their movies but anything that even vaguely looks like something in their movies.
https://www.erlang.org/doc/design_principles/users_guide
There are some good texts that have more examples:
Erlang & OTP in Action - https://www.manning.com/books/erlang-and-otp-in-action
Designing for Scalability with Erlang/OTP - https://www.oreilly.com/library/view/designing-for-scalabili...
One big example of distributed Erlang is Riak:
I suspect that a small team could have most of this
working within the next year or two
Each of steps 2-5 would be an impressive achievement advancing the state-of-the-art. Big "draw the rest of the fucking owl" energy here.* if the platform implodes, clearly the people who were fired didn't build a good platform
* if the platform doesn't implode, clearly the people who were fired weren't needed
Large Language Models have essentially solved language as a problem.
Because garbage like this overselling the tech is how you get another one. Until a week later, when I install updates, and
now half of my applications won’t run because
the new versions are linked against a newer version
of Qt.
"I can't customize it and then expect updates to keep working automatically" != "I can't customize it"The MIT license is for the code artifact; if you want to control the IDEA you'd need much stronger measures like a patent.