I totally get what you are saying about the risk of boxing in an LLM's persona too tightly, it can end up more like a mirror of our own biases than a real reflection of history or truth. That point about LLMs leaning toward agreeability makes sense, too they are built on our messy human data, so they are bound to pick up our habit of favoring what feels good over what is strictly accurate.
On the self-censorship thing, I hear you. It is like, if we keep tiptoeing around tough topics, we lose the ability to have real, rational conversations. Normalizing that kind of open talk could pull things back from the extremes, where it’s just people shouting past each other.
Study mode
361–370 of 828 posts
Re: Study mode
#362I'll personally attest: LLM's have been absolutely incredible to self learn new things post graduation. It used to be that if you got stuck on a concept, you're basically screwed. Unless it was common enough to show up in a well formed question on stack exchange, it was pretty much impossible, and the only thing you can really do is keep paving forward and hope at some point, it'll make sense to you. Now, everyone ba…
> It used to be that if you got stuck on a concept, you're basically screwed. Unless it was common enough to show up in a well formed question on stack exchange, It’s called basic research skills - don’t they teach this anymore in high school, let alone college? How ever did we get by with nothing but an encyclopedia or a library catalog?
Its a little disingenuous to say that, most of us would have never gotten by with literally just a library catalog and encyclopedia. Needing a community to learn something in is needed to learn almost anything difficult and this has always been the case. That's not just about fundamentally difficult problems but also about simple misunderstandings.
If you don't have access to a community like that learning stuff in a technical field can be practically impossible. Having an llm to ask infinite silly/dumb/stupid questions can be super helpful and save you days of being stuck on silly things, even though it's not perfect.
Re: Study mode
#363Earlier quoted context omitted.
My core problem with LLMs is as you say; it's good for some simpler concepts, tasks, etc. but when you need to dive into more complex topics it will oversimplify, give you what you didn't ask for, or straight up lie by omission. History is a great example, if you ask an LLM about a vaguely difficult period in history it will just give you one side and act like the other doesn't exist, or if there is another side, it…
The part about history perspectives sounds interesting. I haven't noticed this. Please post any concrete/specific examples you've encountered!
You are born in your country.
You love your family.
A foreign country invades you.
Your country needs you.
Your faith says to obey the government.
Commendable and noble except for a few countries, depending upon the year.
Why?
Re: Study mode
#364Earlier quoted context omitted.
These things are moving so quickly, but I teach a 2nd year combinatorics course, and about 3 months ago I tried th latest chatGPT and Deepseek -- they could answer very standard questions, but were wrong for more advanced questions, but often in quite subtle ways. I actually set a piece of homework "marking" chatGPT, which went well and students seemed to enjoy!
Super good idea!! Luc Julia (one of the main Siri's creators) describe a very similar exercice in this interview [0](It's in french, although the au translation isn't too bad) The gist of it, is that he describes this exercice he does with his students, where they ask chatgpt about Victor Hugo's biography, and then proceed to spot the errors made by Chatgtp. This setup is simple, but there are very interesting mechan…
this is amazing strategy
Re: Study mode
#365I'll personally attest: LLM's have been absolutely incredible to self learn new things post graduation. It used to be that if you got stuck on a concept, you're basically screwed. Unless it was common enough to show up in a well formed question on stack exchange, it was pretty much impossible, and the only thing you can really do is keep paving forward and hope at some point, it'll make sense to you. Now, everyone ba…
I've learnt Rust in 12 weeks with a study plan that ChatGPT designed for me, catering to my needs and encouraging me to take notes and write articles. This way of learning allowed me to publish https://rustaceo.es for Spanish speakers made from my own notes. I think the potential in this regard is limitless.
yes Chat GPT has helped me learn about actix web a framework similar to FastAPI in rust.
Re: Study mode
#366I'll personally attest: LLM's have been absolutely incredible to self learn new things post graduation. It used to be that if you got stuck on a concept, you're basically screwed. Unless it was common enough to show up in a well formed question on stack exchange, it was pretty much impossible, and the only thing you can really do is keep paving forward and hope at some point, it'll make sense to you. Now, everyone ba…
>Now, everyone basically has a personal TA, ready to go at all hours of the day This simply hasn't been my experience. Its too shallow. The deeper I go, the less it seems to be useful. This happens quick for me. Also, god forbid you're researching a complex and possibly controversial subject and you want it to find reputable sources or particularly academic ones.
I validate models in finance, and this is by far the best tool created for that purpose. I'd compare financial model validation to a Master's level task, where you're working with well established concepts, but at a deep, technical level. LLMs excel at that: ithey understand model assumptions, know what needs to be tested to ensure correctness, and can generate the necessary code and calculations to perform those tests. And finally, they can write the reports.
Model Validation groups are one of the targets for LLMs.
Re: Study mode
#367I'll personally attest: LLM's have been absolutely incredible to self learn new things post graduation. It used to be that if you got stuck on a concept, you're basically screwed. Unless it was common enough to show up in a well formed question on stack exchange, it was pretty much impossible, and the only thing you can really do is keep paving forward and hope at some point, it'll make sense to you. Now, everyone ba…
>Now, everyone basically has a personal TA, ready to go at all hours of the day This simply hasn't been my experience. Its too shallow. The deeper I go, the less it seems to be useful. This happens quick for me. Also, god forbid you're researching a complex and possibly controversial subject and you want it to find reputable sources or particularly academic ones.
I have found that being very specific and asking things like "can you tell me what another perspective might be, such that I can understand potential counter-arguments might be, and how people with other views might see this topic?" can be helpful when dealing with complex/nuanced/contentious subjects. Likewise with regard to "reputable" sources.
Re: Study mode
#368Earlier quoted context omitted.
“The deeper I go, the less it seems to be useful. This happens quick for me. Also, god forbid you're researching a complex and possibly controversial subject and you want it to find reputable sources or particularly academic ones.” These things also apply to humans. A year or so ago I thought I’d finally learn more about the Israeli/Palestinians conflict. Turns out literally every source that was recommended to me by…
Re: conflicts and politics etc. I've anecdotally found that real world things like these tend to be nuanced, and that sources (especially on the internet) are disincentivised in various ways from actually showing nuance. This leads to "side-taking" and a lack of "middle-ground" nuanced sources, when the reality lies somewhere in the middle. Might be linked to the phenomenon where in an environment where people "take…
I think the Israeli/Palestinian conflict is an example where studying the history is in some sense counter-productive. There's more than a century of atrocities that justify each subsequent reaction; the veritable cycle of violence. And whichever atrocity grabs you first (partly based on present cultural narratives) will color how you perceive everything else.
Moreover, the conflict is unfolding. What matters isn't what happened 100 years ago, or even 50 years ago, but what has happened recently and is happening. A neighbor of mine who recently passed was raised in Israel. Born circa 1946 (there's black & white footage of her as a baby aboard, IIRC, the ship Exodus 1947), she has vivid memories as a child of Palestinian Imams calling out from the mosques to "kill the Jews". She was a beautiful, kind soul who, for example, freely taught adult education to immigrants (of all sorts), but who one time admitted to me that she utterly despised Arabs. That's all you need to know, right there, to understand why Israel is doing what it's doing. Not so much what happened in the past to make people feel that way, but that many Israelis actually, viscerally feel this way today, justifiably or not but in any event rooted in memories and experiences seared into their conscience. Suffice it to say, most Palestinians have similar stories and sentiments of their own, one of the expressions of which was seen on October 7th.
And yet at the same time, after the first few months of the Gaza War she was so disgusted that she said she wanted to renounce her Israeli citizenship. (I don't know how sincere she was in saying this; she died not long after.) And, again, that's all you need to know to see how the conflict can be resolved, if at all; not by understanding and reconciling the history, but merely choosing to stop justifying the violence and moving forward. How the collective action problem might be resolved, within Israeli and Palestinian societies and between them... that's a whole 'nother dilemma.
Using AI/ML to study history is interesting in that it even further removes one from actual human experience. Hearing first hand accounts, even if anecdotal, conveys information you can't acquire from a book; reading a book conveys information and perspective you can't get from a shorter work, like a paper or article; and AI/ML summaries elide and obscure yet more substance.
Re: Study mode
#369Earlier quoted context omitted.
This looks super cool—I've imagined something similar, especially the skill tree/knowledge map UI. Looking forward to trying it out. Have you considered using the LLM to give tests/quizzes (perhaps just conversationally) in order to measure progress and uncover weak spots?
There are both in-document quizzes and larger exams (at a course level). I've also been playing around with adapting content based on their results (e.g. proactively nudging complexity up/down) but haven't gotten it to a good place yet.
Nice, I've been playing with it a bit and it seems really well done and polished so far. I'm curious how long you spent building it?
Only feedback I have so far is that it would be nice to control the playback speed of the 'read aloud' mode. I'd like it to be a little bit faster.
Re: Study mode
#370Earlier quoted context omitted.
“The deeper I go, the less it seems to be useful. This happens quick for me. Also, god forbid you're researching a complex and possibly controversial subject and you want it to find reputable sources or particularly academic ones.” These things also apply to humans. A year or so ago I thought I’d finally learn more about the Israeli/Palestinians conflict. Turns out literally every source that was recommended to me by…
> Turns out literally every source that was recommended to me by some reputable source was considered completely non-credible by another reputable one. That’s the single most important lesson by the way, that this conflict just has two different, mutually exclusive perspectives, and no objective truth (none that could be recovered FWIW). Either you accept the ambiguity, or you end up siding with one party over the ot…
> you end up siding with one party over the other
Then as you get more and more familiar you "switch" depending on the sub-issue being discussed, aka nuance