jappgar

I don't know why you wouldn't think they could do this already.

Without the system prompt these models can be used to do all sorts of terrible things.

That's precisely why they need to be strictly regulated by international treaties.

Labelling images, transcription, and vacuuming are things humans can do already.

The vacuum is a great example of an actual invention. Making it "autonomous" just isn't that impressive.

Productivity gains maybe but nothing actually novel.

And those productivity gains are moot if AI costs (including externalities) increase commensurately.

People's workday has been transformed by it. But I fail to see actual transformation beyond "more crap, faster."

AI hasn't done anything we couldn't already do. It's just doing it faster and with more mistakes.

I love AI because it has quickly revealed to the general public how all these tech geniuses are really just rich morons.

It's very much like the Naked Emperor. Sycophantic AI has convinced them their naked inner thoughts are profound and beautiful. Meanwhile the audience is starting to giggle. "This is our CEO??"

My point is that you get used to it. I don't like if aesthetically, but I like that it has an aesthetic at all.

I have coworkers who overuse phrases, oftentimes incorrectly. It doesn't annoy me because I'm familiar with their writing I know what they actually mean.

Because the software platforms are so jank you need dedicated support personnel...

That, and you laid off all your FTEs so you could afford the shiny platform your execs bought after a disgusting night of excess debauchery.

I don't really care if it says load-bearing or belt and suspenders so long as it's using them correctly, which it mostly does.

I don't know how programmers, who are so used to staring at the same handful of keywords every day for decades, have suddenly become so discerning.

Yes, Claude writes boring and predictable prose. It also writes boring and predictable code. That's good!

every one-hit wonder asks the same question.

we liked it until we didn't.

I was involved in both sides of this battle over ten years ago. Things haven't changed all that much.

It's important to note that neither side has moral legitimacy. Not everyone who carries a rifle is a enemy. Not everyone wearing body armor is a saint.

I have given up on the idea that "human vs bot" matters at all when it comes to anything other than voting (which should only be done in person with paper and pen, by the way.)

You could make an argument that "likes" are a form of voting, but you shouldn't. We need to abandon the idea of supposedly democratized algorithms and focus instead on actual democracy.

The losers spending 8hrs of free time on their phone were just watching TV before all this. Bad parents are not a novel invention.

Yes they will approve as many as possible to drive the price of turbines even higher.

Same is true if you're reading a newspaper

It can only deal in tokens, so you're essentially right that it creates a textual description before describing it back to you. This process is obviously incredibly lossy and details are easily missed

I'll see your "government wants biometric surveillance" conspiracy theory and raise you a "pedofiles want to keep kids on social media" theory.

I agree. Very little in here specifically applies to fintech except the ledgering and rounding parts, which are pretty light.

I would prefer to read a defense of something more radical like "database per account." Something that has unique tradeoffs within fintech.

Also, the main advice I would give to fintech engineers/founders is to take risk and compliance seriously from day one.

Financial systems are based around trust. If you don't provably mitigate risks you will lose trust and, eventually, your entire business.

Correct.

Hey don't bring Flash into this

Ironically I don't think AI is this tone deaf.

Is this supposed to be surprising?

refuse to use the s word

strike?

I'm assuming they pay high percentages to processors on deposits to cover the losses.

Then they recoup that (and more) in withdrawal fees

Another reason is that we don't actually know what negative effects regular "scans" might have.

Actually in some cases we do know. Regular xrays are harmful, for example.

And for what? Is it just morbid curiosity or is there something you plan to do with that information.

Today only I'll sell you 6GB of data for 6000USD its a steal. Buy now.

Most requests are reads and letting someone use an invalidated session for reads for 30 seconds on a shortlived token isn't the end of the world, especially considering that the exact invalidation timing and its propagation is already somewhat arbitrary.

For rarer privileged actions you can check a token revocation list.

Julian Assange and Mark Zuckerberg were two nerds on either side of the privacy spectrum in the twenty-teens.

One was framed and tortured, the other was given an empire.

The message was received.

We now only have the Zuckerberg type.

The more guarantees you put in place the more people believe the system is infallible and the more valuable the exploit becomes.

If "signed" photos were treated as incontrovertible truth, then you'll just have people 3d printing hyper realistic masks or something.

It takes considerable energy to train models and run inference. You can't dismiss AI generated content as "low effort", but you can dismiss it as a wasteful diversion.

Robots. That's why they're obsessed with AI and robotics.