Why would you replace an intern with AI? The goal is to give someone experience and make them come back for a job later. If you replace interns with AI, don’t complain about not getting developers later.
Yes, it’s like people complaining that there was a 90% chance of rain and then it didn’t rain.
Of course in reality you either get killed by your partner or you don’t, every judgement except „negligible“ and „extreme risk“ will be wrong if you need absolute certainty. And if you want to have a false negative rate of literally 0, you need to return „extreme risk“ for every single case.
So basically, the sensitivity curves of the receptors in the eye overlap ( https://commons.m.wikimedia.org/wiki/File:Cone-fundamentals-... ) so a signal can’t excite a single type of receptor, but they did it by using a focused laser to exclusively target the one type of receptor?
It is a misconception. Open source doesn’t mean the maintainer needs to interact with you. It just means you can access the code and do your own fork with whatever features you like.
> Ask it to create a Typescript server side hello world.
It produces a JS example.
Well TS is a strict superset of JS so it’s technically correct (which is the best kind of correct) to produce JS when asked for a TS version. So you’re the one that’s wrong.
> Well TS is a strict superset of JS so it’s technically correct (which is the best kind of correct) to produce JS when asked for a TS version. So you’re the one that’s wrong.
Try that one at your next standup and see how it goes over with the team
He's not wrong. If the model doesn't give you what you want, it's a worthless model. If the model is like the genie from the lamp, and gives you a shitty but technically correct answer, it's really bad.
> If the model doesn't give you what you want, it's a worthless model.
Yeah, if you’re into playing stupid mind games while not even being right.
If you stick to just voicing your needs, it’s fine. And I don’t think the TS/JS story shows a lack of reasoning that would be relevant for other use cases.
> Yeah, if you’re into playing stupid mind games while not even being right.
If I ask questions outside of the things I already know about (probably pretty common, right?), it's not playing mind games. It's only a 'gotcha' question with the added context, otherwise it's just someone asking a question and getting back a Monkey's Paw answer: "aha! See, it's technically a subset of TS.."
You might as well give it equal credit for code that doesn't compile correctly, since the author didn't explicitly ask.
As I mentioned TS/JS was only one issue (semantic vs technical definition), the other is that it didn't know to question me, making it's reasoning a waste of time. I could have asked something else ambiguous based the on context, not a TS/JS example, it likely would still not have questioned me.
In contrast if you question a fact, not a solution, I find LLMs are more accurate and will attempt to take you down a notch if you try to prove the fact wrong.
Well yes, but still the name should give it away and you'll be shot during PRs if you submit JS as TS :D
The fact is the training data has confused JS with TS so the LLM can't "get its head" around the semantic, not technical difference.
Also the secondary point wasn't just that it was "incorrect" it's the fact its reasoning was worthless unless it knew who to ask and the right questions to ask.
If somebody says to you something you know is right, is actually wrong, the first thing you ask them is "why do you think that?" not "maybe I should think of this from a new angle, without evidence of what is wrong".
It illustrates lack of critical thinking, and also shows you missed the point of the question. :D
Not really, because humans can form long term memories from conversations, but LLM users aren’t finetuning models after every chat so the model remembers.
He's right, but most people don't have the resources, nor indeed the weights themselves, to keep training the models. But the weights are very much long term memory.
If you want the entire model to remember everything it talked about with every user, sure. But ideally, I would want the model to remember what I told it a few million tokens ago, but not what you told it (because to me, the model should look like my private copy that only talks to me).
ideally, I would want the model to remember what I told it a few million tokens ago
Yes, you can keep finetuning your model on every chat you have with it. You can definitely make it remember everything you have ever said. LLMs are excellent at remembering their training data.
While it's an idealized/toy setting, yes, these are both real categories of sensors. In particular, Sensor B, the "weird one", is just a system that has some defect/failure rate. An example might be a system that only works during the day and fails at night because it uses a camera. Or maybe a camera that's taking pictures from behind a helicopter rotor so it's frequently obstructed. Or maybe you are actually using a bunch of sensors and some of them are broken. (Of course, you'd have to tweak it a bit to get a scenario where every measurement is truly a 50% random split between meaningful and non meaningful, and you can't easily tell the difference, but as I said, this post is an idealized/toy setup.)
You could have some moving element with limit switches and an encoder, that’s pretty common. There’s probably others…
Normally the limit switch would be reliable but it will degrade over time, could be damaged in use, be assembled wrong etc… and the encoder might not be very accurate from the get go.
So if you want a safe/reliable system under as many conditions as possible you might get a problem space like this
That didn’t happen. The person you mean was Buzz Aldrin and it looks like there weren’t even charges filed, and there was no judge involved.
reply