Hacker News new | past | comments | ask | show | jobs | submit | fekunde's comments login

After reading "Why Nations Fail", which I really enjoyed, I wondered whether we could develop an LLM with a wide enough context window where I could input the lessons from such books along with data and metrics about a specific country. Then, the AI could prescribe tangible solutions and policy recommendations tailored to that nation's unique situation. I'm curious what an AI system like that might look like. I also wonder if it's possible to experiment with a smaller-scale version of this concept using today's technology.

I thought the Chicago School had already proved blindly applying "policy recommendations" from on high was a receipt for disaster and misery, but sure, let's do the same with AI, this time it will surely work.

What's the point? We generally know what economic policies work. But they aren't put into effect due to ideology or special interests. This isn't a problem with a technical solution.

Just something I noticed in the Language tokenization section

When referring to itself, it uses the female word in Marathi नमस्कार, माझे नाव जीपीटी-4o आहे| मी एक नवीन प्रकारची भाषा मॉडेल आहे| तुम्हाला भेटून आनंद झाला!

and Male word in Hindi नमस्ते, मेरा नाम जीपीटी-4o है। मैं एक नए प्रकार का भाषा मॉडल हूँ। आपसे मिलकर अच्छा लगा!


In my language (Romanian) it refers to itself as a male. Although if I address it as a female it responds as one.


Yeah, the pricing tiers make it all the more morbid and disconcerting. Scary future for sure.


2211 1122 1089; 9801 1089 8712; 8721 1278 7443; 7443 3447 3996; 9963 3699 6264; 6642 2466 4176; 7641 1467 6174


This is so great for mockups where you have a faint idea of what you want [0] Treasure map - https://scribblediffusion.com/scribbles/4unwo6iq2jhwlmuqpu6v...

[1] “treasure map in the style of lord of the rings maps” https://scribblediffusion.com/scribbles/t36as45npjapxfekcuuy...


I was looking for this comment. It is one of my favorite takes on how a simple program evolves to become the all controlling entity.


This story has stuck with me since I read it many years back. And then seeing it come to life in so many ways where machines are "optimizing" processes and workflows for humans is scary.


HackerRank's practice questions are pretty good.

https://www.hackerrank.com/domains/sql


Any anecdotes as to how they do it?


Hard to say any specific anecdotes for how he does it. It's mostly in attitude; how he says things rather than what he says. You can tell from his attitude that when he challenges you, it's not to put you down or exert authority, it's because he genuinely wants everyone to come to the strongest conclusion. (Even if most people think they also act this way, probably they don't, or they don't express it that well).

There will be a discussion that starts meandering, going down rabbit holes, and somehow he'll find the perfect question to ask, which is always a very simple one, to set everyone back on target. Sometimes I don't think people even realize this is happening.

The reason it's hard to explain is because it comes across as an innate character trait, a skill that's just core to his personality, which is why it's so hard to emulate.


This looks like the way to go for people like me who have used everything from uninstalling apps (then resorting to the browser) to installing usage trackers. Just eliminating the ability to do anything but the minimal.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: