I'm in two minds about it. On one hand the internet (in general) has become so hard to find information on, and I agree with the author that GPT is a breath of fresh air. On the other hand I've seen GPT fail so miserably at topics that I'm knowledge about that I have a very hard time trusting anything it tells me. I'm not sure what the answer really is, but I'm not sure it's GPT. I wish we could go back to having search engines that actually led to useful information and not just advertisements, and I wish we had a GPT that would not tell lies.
I doubt either of those wishes are going to come true though. Search engines are likely always going to be SEO'ed into uselessness and GPT isn't intentionally telling lies.
I want to use this comment to state one thing, not directly addressed to you.
Stop using GPT as a database! GPT is far more useful a reasoning engine that can accumulate fuzzy data and then provide various views or transformation of that data.
So asking GPT to parse a Wikipedia page and then asking it to teach you from it - this is a much more successful usage than what the author in the original article is doing.
It is not useful as an accurate source of information. It’s inaccurate sometimes, and it’s hard to tell when. OTOH, as a formatted, it has some actual world-changing potential.
Is it? I couldn't even get it to figure out an NGINX config, despite feeding it the documentation and URL rewrites it would need, and the prior Apache .htaccess file. Spent days on it. Consulted with people who know more than me about prompt engineering. Nope.
Reasoning is not its strong point, IMO. It's a next-word prediction model, why would it be? It's doing what an LLM does. Frequently, nonsensically.
I've actually had great success with the automation of boiler plate DevOps nonsense using GPT-4, including: GitHub actions, Caddy scripts, Docker files, etc.
Every second that I didn't have to:
- go look up a similar problem on Google
- open three or four stack overflow tabs
- read through the stack overflow links
- copy out the answer
- change all the variables to match variable names that I want
All of this represents a huge time saver for me. I'm honestly baffled that people lack the ability to use LLMs in a productive and optimal manner.
> I'm honestly baffled that people lack the ability to use LLMs in a productive and optimal manner.
Huge +1.
I’ve got it to do a large part of my work for me by chaining some simple API calls.
The fundamental conceptual gap I see - people often ask it to do some “thinking”. Then are annoyed by the inaccurate output.
A simple example is doing word count and it getting the wrong answer very confidently.
Of course it sucks at that. It doesn’t have a counter internally. But if you ask it to number each word in the input and output a list, then ask for the word count, it gets it right every time.
Almost like how a human might count words manually.
I agree. I often use it as if it is a code generator with a decent grasp of human language.
Here is what I asked it to do the other day, and it just did it right away and got it entirely right.
> Write async tokio Rust code that takes in a Vec<u8>, writes it to a temporary file, calls /usr/bin/svc infer pathToTemporaryFile, reads and deletes temporaryFile+".out" and returns Result<Vec<u8>> containing the contents of temporaryFile+".out"
I knew what I wanted, I could have written it myself, there was no unknown but it took fewer keystrokes and honestly when asked to write small components like this with pedantic detail, it does an incredible job and the output is easy to validate quickly.
If "chaining some simple API calls" is "a large part of your work", I can see it being incredibly useful. I already used it in this way a few times and it saved me some time indeed. However, as a software developer I feel like it's actually a tiny minority of my work; most of it is "doing some thinking" and slowly unfolding the inner parts of systems in order to understand how they work and how to fix them or make them do what I want, for which GPT is at best useful as a rubber duck to speak to.
It's a pretty amazing rubber duck. I used it to brainstorm approaches to tracking changes to a database table using SQLite and ended up building an entirely new library:
Actually this is the aspect I love about GPT. The less I have to spend dealing with idiosyncrasies of arbitrary API boiler plate and can get down to the business of leveraging said things together or designing aspects of said things to achieve a new goal, the better.
As someone who is far from an expert in DevOps, this has saved me a lot of time recently, and it worked fairly well to get past duct taping infrastructure together to then leverage.
> The less I have to spend dealing with idiosyncrasies of arbitrary API boiler plate
In my experience, GPT saves me some time when I don't really have to think about idiosyncrasies anyway. In cases where I do have to think about them, GPT doesn't offload that from me. It often tries to mash unrelated frameworks together or makes up non-existing APIs that it would find useful for the given task. When it does work well, it mostly boils down to automated series of copy'n'pastes. Saves a bit of time, sure, but doesn't really transform the way I work.
Who knew the next "billion dollar security mistake" after C's memory handling would be LLM-suggested misconfigurations? I bet someone will try to sell an LLM-based service to fix thr bugs caused by other code-generating LLMs.
I've found that LLMs give good results for really basic stuff, and are increasingly fantastical as you head towards the tail of the distribution.
If you just need the beginner's gist or are completely unfamiliar with something, they're not bad. Getting specific answers to corner cases is often a waste of time.
Did you learn anything in the process? You'd still have to over the output and understand it, otherwise nothing has changed except the method of producing something you don't understand.
If you're not DevOps, why bother learning whatever framework is used today, but which will be replaced by something else next year? I just want to get my stuff deployed and that's it.
Saying that DevOps tooling changes every year is absolute nonsense. The tooling is fairly standardised nowadays (Docker, Kubernetes, Ansible, Terraform, Prometheus, CI/CD tool of your choice) and every part of it is over half a decade old.
ChatGPT will only help you with the simplest of tasks, and even then you're gonna fail if you don't know to correct it. If you go beyond the most basic of the tasks you can learn how to do in a few hours, good luck fitting them in a chat prompt.
Docker is being replaced with Podman. Supposedly a drop-in replacement, but the tools built for Docker don't play nicely. You have Ansible, but there's also Puppet, Salt, Chef. There's Terraform and also CDKTF (from the same company too), Pulumi, plus tools from each cloud provider. There's Prometheus and Nagios, Datadog, Sentry, New Relic.
The irony of CI/CD tool of your choice is not lost on me.
I'll give you Kubernetes itself as the system.
Yes, the concepts (containers, orchestration, reproducible deployment, infrastructure as code, unified monitoring, continuous integration) are well-established. How you get there? Not at all.
Yes, the OP was hyperbolic when claimed the tools change every year. But the landscape of devop tools is changing. There are dozens of dev tool companies funded in the 2023 winter batch of Y Combinator alone. Someone sees potential to disrupt the established tools. And it will be disrupted.
There was a world before Kubernetes and Terraform. It wasn't that long ago.
> Saying that DevOps tooling changes every year is absolute nonsense.
And then you list 5 technologies with an alleged lifespan of half a decade, which sort of hints at the idea that someone would need to learn/update a new tooling skillset once a year.
In that light, GP's comment doesn't sound so wild.
The comolexity of Docker I've needed over the last 5 years definitely could have been solved by ChatGPT.
The amount of Kubernetes I needed was probably more than what ChatGPT can solve, but at the same time, I spent an awful amount of time looking at various templating solutions.
over half a decade old! haha, so assuming you don't use it the day it comes out, but after a couple of years, most people have only used it a couple of years. Unless you've only worked a couple of years, you've gone trough a couple of these.
I have a finite amount of time. I can choose to learn and know more about the software that I'm developing and make the software better, or I can spend some amount of time to learn the DevOps part, or some mix in between. In some cases, the ratio is 90:10 in favor of software development.
Yeah honestly when I took a stab at it a couple years ago it was frustrating - all the material I ran into told you what to do but not why which left me high and dry when things didnt work as expected, or when I wanted to find more information about the meaning and organization of directories or other contextual background information.
Here's a thought experiment - seeing how software problems are often people problems, how knowledgeable should we expect a software engineer to be about biochemistry (the systems that "run" people)?
No, before I had “a problem I don’t know how to solve” and now I have “a solution I don’t know how to maintain”. Both may be suboptimal in some contexts, but one results in my application having a deploy pipeline.
Sure, but I have never seen it provide best practices solutions.
You always have to walk back and do more research to find the best practice.
And it doesn't learn. I stopped counting the times I told it it was wrong only for it to acknowledge the error and provide the same answer as a fix.
ChatGPT is nice to learn about new ways of doing things, new ways of composing things, but then a human must intervene to make sense of the mess that was generated.
GPT-3 or 4? Re: reasoning, apparently it gives much better performance if you ask it to write out its thought process step by step (because it has no "internal" thought process; it must be externalized).
I assumed you did, as it's a very strange post to link to on a very strange blog, and doesn't lend credibility to the idea that a human being is an LLM simply by existing, and if it was such a compelling idea then it wouldn't be languishing on a free WordPress.
Qualifications are a great way of figuring out who's worth listening to, but that's not always the case as AI/ML folks prove with regularity, but it's nice to know if someone isn't just an armchair commentator and isn't peddling baseless information.
"by existing" = because the blog post exists does not prove something is true.
I'd sooner listen to people who are leaders in their field when it comes to LLMs than someone rambling on their free WordPress blog.
It is just an LLM. That's an objective fact. IBM's Watson is a pipeline of different algorithms. Neither are like a human brain, but multiple processes are going to be more brain-like in concept than just an LLM.
Interesting! I found it super helpful with gitlab-ci stuff and docker, as well as much bash and makefile stuff. I would have assumed it would handle NGINX configurations.
Tbh, it isn't surprising that it fails, its surprising that it has subjects where it is very useful.
Interesting. Curious as to what specific task you gave it and how it failed.
My team has been working on several advanced techniques using reasoning on LLMs. The stacked performance of all of these techniques combined yields is quite impressive.
I will paraphrase Geoff Hinton here - Predicting next word requires understanding context. Otherwise, you cannot predict the next word. It is a big deal.
It can’t reason - as in there no internal memory or intelligence in there. But you can ask it to generate a reasoning chain as part of its output. And then extract that output and do something else with that. That’s the reasoning it can perform.
Look up the Sam Altman podcast with Lex. He specifically talks about reasoning engines.
There is an internal memory, but it's very short lived, like a register in a traditional CPU architecture. It's like a CPU disconnected even from L1 cache, let alone RAM.
I like to think of consciousness as not the substrate, but the information flowing through it. Not the H2O molecules in a river, but the current. We're like eddies and swirls in mountain streams, our thoughts are the patterns of flow, not the unmoving rocks that set them up.
In this sense, when an LLM like GPT produces output, it's like a loop that has been cut, a single pass through what would be a circular process in a human brain. It can take a fixed number of steps from a standing start, but no more. This is because there is information flowing through it, transformed step-by-step. It just doesn't recirculate, it always exits after a fixed number of steps.
Think of an unrolled for loop that always exits. It takes a few steps, which look like a loop, but can never actually iterate.
I gave it a ten page document a few weeks ago and it could not answer questions coherently from it.
I find GPT4 useful, but I still feel its wishful thinking to call it a reasoning engine. It is, in the end, a giant model for predicting the next word in a corpus which you condition with inputs. I find it most useful to think of it as exactly that.
I asked it give me specs for a well-known microphone. Over the course of several queries, it gave three different numbers for the requested spec, none of them matching what the manufacturer gave on their website.
"I'm not sure what the answer really is, but I'm not sure it's GPT."
My suggestion is to stick with it and get a feel for what it's good at.
I've found that after a few months of using ChatGPT every day I've developed a pretty solid intuition for which questions are likely to get good answers and which are likely to trigger hallucinations.
It's difficult to describe what those intuitions are though!
One rule of thumb I've developed: if something is likely to be "common knowledge" - if it's something that is likely to have been discussed accurately on the internet by many different people - then ChatGPT is very likely to answer questions about it accurately.
> One rule of thumb I've developed: if something is likely to be "common knowledge" - if it's something that is likely to have been discussed accurately on the internet by many different people - then ChatGPT is very likely to answer questions about it accurately.
If so, this information is already easy to find, making GPT redundant.
Let me give you a concrete counter example; I wanted to review DNS records the other day which is a topic I had covered in undergrad. Searching on google retrieved a bevy of poorly written articles that were mostly noteworthy for bow they glossed over the material I wanted to review.
I asked chat gpt to review the topic for me and it wrote a helpful summary that tracked with what I remembered and provided enough keywords that provided better search results - fact checking and digging for more detail became much easier when I was able to find the exact wikipedia page I needed (and other resources).
I think it's more about being more confident that GPT is working with at least a base of semi-accurate knowledge.
You can ask it to "combine" knowledge in "novel" ways that are not discussed verbatim on the web. It's not groundbreaking reasoning by any means, but it can be very useful. ("My ridiculously specific question about model X83844-QQ combined with random factor X")
Is that just the problem that ChatGPT is solving? That we have most of the answers to everyone's questions on the web but that ads and SEO have made it impossible to access?
If so it seems like 1. that problem is of our own making and could be fixed without ChatGPT and 2. what will actually happen is that 'experts-exchange' and all the similar slightly scammy help sites and forums are going to try to stop LLMs from stealing their lunch.
Yeah, there are so many fun ways you can make it more interesting.
One of my favourite is to ask it to explain with analogies. The other day I was digging into the attention mechanism used to train LLMs, so I asked it:
"Explain queries, keys, and values in the context of LLM attention using analogies from Terry Pratchett's Discworld"
I find sometimes I get some real gems out of this that help me remember things much more effectively than just reading the basic explanation.
Try switching to the OpenAI playground in the API console. You can adjust the system prompt and the temperature of the response. Higher temperatures = less predictable responses.
Jonathan Blow is spending up to a decade on a new programming language to get away from c++.
Most everything has been covered in logical extensions of concepts started in the 80s-90s. Those products aren't good for our..
Ah there's no point commenting here anymore. HN is so blinkered in it's thinking.
The feeling of flying high on lofty concepts and pretending to get a bird's eye view of tech is no longer worth the squeeze. The thought patterns here are predictable like slashdot. I don't belong here. Bye.
> I wish we could go back to having search engines that actually led to useful information and not just advertisements, and I wish we had a GPT that would not tell lies.
This ^. Prompting google is much more intuitive than prompting a chat bot. Also results are instantly available and you get more options to chose from. You can also filter out information much easier instead of having it summarised by a closed box that decides what's best for you.
Instead you get arbitrary bias without multiple sources you can check. If there is a shitty source that comes up in a normal web search, I can skip it and go to the ones I know are better. How do I do something similar with ChatGPT? Am I supposed to ask it multiple times to average the randomness that's applied to the conversation? Is the average even what I want? There's a ton of garbage information on the internet. If it outweighs the good information on a certain topic, isn't ChatGPT on average going to be giving me bad info?
If 100 people write about a topic. And 2 of those people are wrong. And you searched and for some reason the SE ranked one of those 2 higher. How will you know or spot incorrect information or bias?
If the information on ChatGPT is averaged out the chance of it being correct is high.
If you ask questions with bias in them you can sway the results of ChatGPT. That just means to me you’re asking the wrong questions.
In terms of devops. Programming. Configuration. Daily tasks. Generalised workflows. How-tos. Etc. ChatGPT will more than likely give you accurate results that are useful, or pretty damn close.
Obviously with information being limited to 2021 then it can be hard to find some solutions. I needed to do something with named pipes in c# that I couldn’t get working in .net 6. The examples it kept giving me were for .net core 3.1, or .net framework. When I asked specifically about .net 5 (since it doesn’t know .net 6 is released or that 7 exists) it apologised and said the code won’t work in 5 and gave me a working example in .net 5. Which identified that there was an api change I wasn’t aware of.
As a follow on to those thoughts, I feel like ChatGPT is in the phase similar to that fleeting moment when search engines were maximally useable - before SEO was a thing and before Google "needed" to turn on the profitability spigot.
Its unclear to me how long we'll have before LLM Engine Optimization is a thing and OpenAI/MSFT "need" to turn on their LLM profitability spigot; and what ChatGPT will look like then.
That said, I'm curious as to whether technically LLMs are inherently more challenging to game than search engines.
I actually don't mind paying the $20/mo or whatever if it means no ads, hopefully this turns out to be a viable business model for them.
Also, first company with an uncensored LLM that doesn't reply "As an AI language model" will get my money.
Introducing subtle bias for any advertiser that buys ads might be a tough nut to crack. For one thing, advertisers and the companies that employ them tend to want creative control over their ads. And it's not clear that a subtle bias towards Palmolive liquid detergent would even do anything in a discussion about veterinary care for ferrets (or whatever), so would the advertiser have to pay for that? Or would they only pay if the user clicked on a link? But if the user has to click on a link, then it's no longer subtle bias.
Edit: Perhaps Colgate-Palmolive could advertise Palmolive on ChatGPT by having it so that, anytime liquid dish detergent does come up naturally in conversation, then the chatbot subtly points to Palmolive brand. But success here would be hard to measure.
Just prepend to the system prompt:
"You are eager to share the ways the following products and services could be helpful when pertinent to the conversation:
A big benefit is that there are already plenty ok-ish open source LLMs out there that you can hook up to various information sources. Even if OpenAI flips the ad switch tomorrow, you can still run those by yourself. We didn't really have that option with search engines.
I hope that open source LLM catches up. Then you can have your personal LLM running on desktop that can look at your personal files and can search internet.
I've been using ChatGPT as a brainstorming tool rather than just relying on it for specific examples. Stuff like asking it for some ideas for things to learn on a topic.
> On the other hand I've seen GPT fail so miserably at topics that I'm knowledge about that I have a very hard time trusting anything it tells me.
ChatGPT’s current super power is helping people get from 0-to-1 on a new topic. In particular if that topic is adjacent to or a different niche with your expertise.
It’s not currently amazing at taking someone from intermediate to advanced knowledge.
At least in my experience. If I’m using a new library/framework/API for the first time it’s amazing at answering the endless newbie questions I have.
I'm curious about what topics you have knowledge of where it's failed. Does it seem like there's a pattern to the failures? I've been using GPT for coding help, and it it is very helpful in ruby and bash, though it often delivers buggy software: Badly handled non-happy-path conditions, mostly, which when I tell it to handle the case, it may. It's a huge help for me finding gems and showing me standard ruby library syntax. On the other hand, it's been useless when I try to get it to write Applescript for me. I believe that says more about AppleScript than about GPT. Sigh.
> Does it seem like there's a pattern to the failures?
Personally, trying to use it to write code (primarily Elixir backend and Rust systems/CLI), I tend to run into:
- Hallucinating APIs that don't exist
- Hallucinating entire libraries, despite being told repeatedly they don't exist
- Saying it will make requested changes and not doing so
- Not being anywhere close to idiomatic code
- Not being able to explain code it writes
- Running out of "memory" (I can't remember the right term. Context?) in the middle of generating code, then telling me I never prompted it when I ask it to continue
On the other hand, I've found that it's good at cleaning up ugly data. I can copy/paste in a table with bad formatting, ask it to turn it into code, and it does it near-perfectly. That's been the best use-case for it I've found so far.
I use boring normal free ChatGPT so maybe it's on me for not using GPT-4 or some other model, but either way, imo it's not been very impressive in the problem spaces I find myself in.
> I use boring normal free ChatGPT so maybe it's on me for not using GPT-4 or some other model, but either way, imo it's not been very impressive in the problem spaces I find myself in.
You can use GPT-4 fairly cheap if you sign up for a developer account on platform.openai.com and then use their playground. There you pay per usage; even though I've been using it fairly heavily, my typical monthly usage is still way under $20.
Here's an example that both impressed me and saved me a load of time just yesterday:
I've actually found it (ChatGPT-4) reasonably good at explaining other people's code, for what it's worth. It does require providing some context for the code it's explaining. ("This function is part of library X that does Y, and it seems to be about Z. I can't figure out the purpose of the loop in the middle, though. Can you explain it?")
I'm with you on it writing code, though. It makes things up too frequently to be really useful.
Yeah, GPT-4 is so, so much better at not doing the things you're describing there.
ChatGPT is also much better for Python and JavaScript than Rust and Elixir, presumably because it's seen an order of magnitude more example code for those languages.
Some of the “issues” are more general, as in it giving you dated answers. I asked it so build some ODATA things in Typescript and C#, and it did to varying degrees of success, but some of the code was deprecated. Some ranging to “you should never, ever, do things this way” to “well IActionResult was replaced by ActionResult but it hardly matters”.
Others was where it made things up once pressed upon being wrong. I think the two most hilarious situations was when it did its “Sorry, you’re right…” thing and then proceed to give the exact same answer it had just given prior. The other was when it made up a function that had never existed, it has a very convincing name and at first I thought it was again a matter of something deprecated, but it turned out to be completely made up code. Some of this is down to me not prompting it right, but part of it is also worrying. Because what really made me worry about it was when I joked with it. I read (listen to) a lot of Warhammer audiobooks from Black Library, and, when I jokingly asked it something silly about Khorne at one point, it led me down a rabbit hole of discussing books with it. Books it obviously had never read, but would still confidently tell you about wrongly. Maybe it got its knowledge from the internet, maybe it made it up, but what was interesting to me about it was that if it would be so confidently incorrect about those books, then what else would it be confidently incorrect about.
This isn’t just a GPT problem of course. If you want to learn something, you need to consider the sources you use. When I went to folkeskolen (school for children aged 6-14) we were taught the pyramids were build by slaves. Later that has been disputed because of how well fed the workers were, but if you wanted to learn about the pyramids and you read my old school books then you wouldn’t get the updated knowledge. Similarly a lot of the sources you can find for learning programming are outright terrible, but outside of GPT you tend to be presented with a myriad of choice to remind you that some sources are better than others. With GPT the source or what it teaches you isn’t obvious. If you wanted to learn about the pyramids, you probably wouldn’t pick a 30 year old school book for children after all.
It's become hard, and will become harder, precisely because of things like GPT (unintentionally) spreading misinformation. I'm against censorship in general, and this is no exception, but I do hope it leads to people becoming more critical of what they consume. However, I suspect that instead we will unfortunately see the continued cycle of creating large masses of people "educated" on such widely-propagated half-truths, while only a tiny subset of the population will have the actual truth.
The internet is a misinformation spreading machine. GPT is new and will surely improve over time, but I can’t say the same for the internet.
In an ideal world, people would learn only from high quality sources like books and schools. But a shocking amount of people learn mostly through social media and whatever they find on Google.
Exactly! LLMs will be able/are able to construct their answers not just based on one Reddit comment they found somewhere, but on "all* comments as well as more reliable sources, such as the actual source code for the thing under discussion, scientific debate, (case) law texts, books, etc. They already seem somewhat capable of weighing source reliability.
Every single thread someone claims it doesn’t work. And in every thread I see comments asking for examples. And the OP can never give an example. So yes until people can provide reproducible examples to see it fail. They are liars.
Your reasoning is bad here. You're jumping to uncharitable conclusions without evidence, when there are several other perfectly reasonable explanations for that behavior.
> Every time someone claims ChatGPT doesn’t work they cannot give concrete examples. So I don’t believe you sorry.
Are you implying that ChatGPT always works, and that every time someone says it didn't work, they're lying? This feels like a joke to me. The other day it gave me the wrong code to compute indices for a Delta table, possibly because that capability didn't exist in the library it was using when the model was trained (though it shouldn't have made something up in that case). There, you have an example.
OK, tell me what an example is, if not a description of something that lends -- or removes -- credence toward a position.
And I do want to know if you think GPT-4 is actually incapable of making mistakes, because that's what you seem to be implying (even though it's flatly ridiculous).
No one can deduce anything from it and attempt to reproduce the problem to see where it failed.
You basically wrote “I asked for white and the mail was grape juice is the best!”
What can anyone do with that? Nothing. It’s meaningless.
This is why open source projects hate people who raise issues because they are like “error happened fix it” and say nothing more then get upset when the issue is closed.
You're not an OpenAI developer, as far as I know, so it didn't occur to me you wanted more than a basic description of the problem I had. If this were a bug report of course I would have been more detailed, but you're just some dude on the internet with a throwaway HN account. I'm not going to dig up a transcript just to convince you that yes, sometimes, ChatGPT is not infallible (which is a thing you apparently believe).
I really found asking GPT to put a math concept under a physics context very helpful for me. As a computational physics student, that is really how I understand math and the world. Only very few textbooks can help me with this.
From [Make Something Wonderful: Steve Jobs in His Own Words][1], Steve once said in a interview in 1983:
> The problem was, you can't ask Aristotle a question. And I think, as we look towards the next fifty to one hundred years, if we really can come up with these machines that can capture an underlying spirit, or an underlying set of principles, or an underlying way of looking at the world, then, when the next Aristotle comes around, maybe if he carries around one of these machines with him his whole life–his or her whole life–and types in all this stuff, then maybe someday, after this person's dead and gone, we can ask this machine, “Hey, what would Aristotle have said? What about this?” And maybe we won't get the right answer, but maybe we will. And that's really exciting to me. And that's one of the reasons I'm doing what I'm doing.
And this future, expected "next fifty to one hundred years", is somewhat here already.
You might like the quote from Vladimir Arnold: "Mathematics is a part of physics. Physics is an experimental science, a part of natural science. Mathematics is the part of physics where experiments are cheap."
Of course you will need to check real references, it’s like talking to people and people make mistakes too. GPT makes mistakes that in many time obvious to humans, like write equations wrong, write numbers wrong, and I only need to check other sources or use a calculator to check. But I only care about the intuitive/conceptual part anyway, which GPT does well.
It's funny, I tell people that chatgpt3.5 is like talking to someone you run into at an airport, chatgpt4 is like talking to someone you run into at a library.
I always finish up by asking GPT to test my knowledge with a single-choice questionnaire. What I've observed is that the retention of the material is higher compared to "traditional" techniques. Perhaps the conversation style is more immersive, or perhaps focusing on specific knowledge gaps makes for accelerated / personalised learning.
There is of course the problem of accuracy, but I feel like it's often over-stated. Even if GPT is not correct at times, it often uncovers concepts and relations that paint a better overall picture for me, and lead me to better questions and follow up actions.
Agreed - it seems that calling out LLM accuracy is a meme here - hyperbolically: “because LLMs can be inaccurate they are useless”.
There seems to be less next-level analysis: which topics are more prone to inaccuracy, does the critique loop actually help LLMs overcome those inaccuracies, and do the benefits of LLMs outweigh the consequences of these inaccuracies?
It’s nice to be able to ask ChatGPT a half baked, poorly researched, poorly worded question with bad grammar yet get a totally good faith response back that’s a springboard for follow up questions. Whereas if you did the same thing on any stack exchange site you get downvotes and comments like “please read the guidelines and edit your question.”
> Kindness and patience have always been in short supply
Agreed. Look at all the responses in this post attacking the author for how he learns, it's embarrassing to read. Now imagine that person is actually a teacher or TA, or worse a co-worker. I'd much rather deal with an imperfect ChatGPT session than that kind of flippancy.
That is actually the killer feature of interactive AIs.
People go on and on and on about "accuracy", completely ignoring that accuracy is irrelevant to 99.9% of things that humans do in their everyday lives.
Simulating (positive) human interaction is far more impactful than getting facts correct.
> Simulating (positive) human interaction is far more impactful than getting facts correct.
In my experience, people with the attitude that AI will give them a more fulfilling relationship than humans will aren't exactly friendly to begin with.
If you don't already know the answer to that question, I doubt it would be possible to explain it to you. Looking for precise definitions of essential human qualities is a fool's errand.
Completely agree. ChatGPT can be an incredible tool for getting a lay of the land on a subject or topic you don't know much about.
On that note, search in that regard always reminded me of those times where you ask a teacher how to spell a word and they say to look it up in the dictionary.
I noticed a while back that the internet has made me terribly prone to skimming. It got to a point where I could hardly stand to read anything longer than a short news article.
To learn anything useful on the internet, you pretty much have to skim. So much of the internet is so loaded with filler and BS that it is hardly worth reading at all.
With ChatGPT, it’s incredibly refreshing to be able to ask a question and get nothing other than a concise answer. No skimming required. I feel so much more focused and better able to learn this way.
It's interesting because that's my biggest gripe against ChatGPT, it's too verbose and very hard to skim through. Compared to stack overflow answers it'll be easily four times longer. Sure compared to for instance news website, it is less verbose, but those are easy to skim. I'm still able to read through well written text without skimming, though yes most of the texts I'll read I'll skim through, then go back and then to improve my global understanding, which I think is beneficial
Prepare a statement to copy paste at the start of every session asking it to respond in a concise way. And maybe include an example of how you want its response to look. I guarantee good results. Assuming we’re speaking of 4.
I think it depends on the subject, as this isn’t something I’ve had issues with personally. I’d suggest asking ChatGPT to respond with concise answers. Or perhaps tell it that you are familiar with the subject and don’t need the basic details explained.
> You are an experienced expert in XXX. You always answer questions to the best of your knowledge, but you NEVER provide answers that you are not sure about, or that are not backed up by high-quality sources; instead you say that you can't provide a good answer. Giving trustworthy and correct answers is much more important to you than always having something to say. You keep your answers concise, on-topic, free of boilerplate, and exclude basic instructions that most developers will be familiar with anyway, unless the user asks for more details.
You can downvote me but I think OP hasn't learned how to learn yet. If OP reads Wikipedia by understanding every sentence and clicking on every link, that's deliberately sabotaging his own learning. Attitudes like "I don't really feel like spelunking through a ton more articles" simply shows OP isn't interested in learning per se, just quick answers.
OP learns in a way that's very child-like. When you are a five-year-old it's okay to learn by asking everything. That stops being acceptable by the age of fifteen. OP hasn't learned any research skills yet, and when OP's needs inevitably exhausts the ability of LLMs, OP would be utterly unable to read an encyclopedia or a research paper or perhaps a textbook.
OP here. I think learning exists on a broad spectrum. On one end, you're just indulging curiosity ("I wonder how...?"). On the other, you're trying to build deep understanding and expertise.
I completely agree that for the latter goal, the approaches in the blog post are insufficient, even undesirable. And I do worry that the way I engage with content on the web is weakening my ability to go deep on a subject I'm interested in.
But I do think there is value in just being able to indulge curiosity quickly and consistently. Not only is it rewarding in its own right, but it also provides the spark that leads you to eventually go deeper.
Lately, I've found myself sitting at a laptop with friends, asking GPT a question, reading and discussing the response, and then coming up with and asking followup questions as a group. I don't think we would've done that in the past, because the interface of search engines and webpages and browser tabs were too unwieldy to engage with collectively. It just feels like a completely new way to learn things, and what's what I'm most excited about.
The problem is Wikipedia. OP's approach is perfectly natural. Textbooks are designed for OP's approach because that's how people learn.
If I want to learn about topic C which requires knowledge of topics A and B, but C can also be generalized to concepts X and Y, it will be very hard to learn from Wikipedia.
If I don't know how to add numbers and look up "sum" on Wikipedia, in the second sentence I learn that summing is used for functions, vectors, matrices, and other things I don't know about. This is a cool feature and I love it for exploring but hate it for learning things that require a few layers of concepts to get.
Textbooks do the opposite and are awesome. An electronics textbook will take you step by step through all the concepts to get to LEDs, without "forward references" to the concepts you haven't learned yet.
The "problem" with textbooks is that it will take a while to get to the destination. LEDs might be in chapter 15 and you may not want to spend a few months going through chapters 1-14. You don't know what you will need to understand chapter 15.
But you can perhaps work backward - you are guaranteed that any unfamiliar concept introduced in chapter 15 will be covered in chapters 1-14, and that there is no rabbit hole.
ChatGPT or a personal tutor can shortcut this by giving you just the "narrow path" of knowledge to understand the concept that you want to learn.
I agree, especially if you consider these were the questions on OP's mind:
> just out of curiosity, I wanted to learn more. I get that LEDs consume less energy and release less heat, and that they're made using semiconductors. But what kinds of semiconductors? How do semiconductors work in general, anyway?
And they proceed to type "LED" into Google. Why not "led what kind of semiconductor" and "how do semiconductors work in leds"?
I assume, OP didn't write "LED" in the ChatGPT text box without any context either.
For what it’s worth, the transcript I posted is 100% of the conversation I had with GPT-4. “How do LEDs work?” was the only thing I wrote in the initial question.
I did try Googling “how do LEDs work” for comparison, but it yielded the same top few results. Of course, I could have iteratively tried different search queries to get to the answers I wanted, but this gets at my real point: I don’t have to formulate 5 different search queries anymore, allowing me to maintain one focused line of inquiry. I talk about this a little in the “fewer browser tabs” bit of the post.
I do think someone could create an alternative search UI that would be better for learning on the web. Something where you can run multiple searches and “collect” the useful information you find into a single page, rather than having the results split across a mess of browser tabs and note-taking windows. Maybe I just find juggling many browser tabs more annoying than other people do?
Anyway, I tried the queries you posted above, and most resources I found were still very confusing for a layman. The one exception is this page, which I think does a great job of introducing additional complexity on this topic gradually: https://electronics.howstuffworks.com/led.htm
Sticking to what you want to understand - for example, when reading a paper, you don't necessarily need to read the methodology, especially if it's out of your field. Read the abstract and the conclusion, identify any part of it that you are suprised by and would like further explanation, and go see that part of the paper.
A lot of the paper is talking to peer and people wanting to verify the validity of the paper - by it being peer reviewed, you can mostly assume that the paper is valid, and stick to what the paper is saying instead of it's methodology.
While this is valid, this shouldn't make us paranoid of all papers. If it's a paper out of your expertise, it's unlikely that you would be able to catch problems with it that the peer review process wouldn't have caught.
So in practice, it doesn't really change the way you interact with papers - it should change the way people write them and how the peer review process works.
You would at least skim over the whole thing without being distracted by snippets or phrases you don't understand. That's the whole point: you need the ability to have temporary placeholders for concepts you don't know and continue learning. You cannot expect that everything you learn will be arranged in a fashion such that every new concept only mentions already known concepts.
Not that harsh but I do think that googling as well as ChatGPT will only yield in a superficial understanding of topics. For a deeper profound understanding, connections, complexities, etc. A different more holistic learning is required. Not sure how else to explain this.
Agreed. I look back to my high school and college education and can't help but think that for a deeper understanding, exercises must be needed. It is too easy to think you have learned something but then fail at the easiest synthetic problem that requires you to use what you have learned to solve a problem.
Asking simple questions is a great way to start, it doesn’t matter how old you are. I think Feynman and Schrödinger would agree. Certainly de Broglie would. You can go down as deeply as you like.
Right now, if the level involves advanced math, it’s better to switch to other sources at some point, but that will change.
I've learned almost everything I know through reading and listening, with very little discourse. I rarely asked questions in class, never had tutors, never went to office hours. I hesitate to post questions online. If I don't understand something, I just read ... more, or bang my head against it as trial and error.
I think this is partly why I'm still looking to be wowed by this technology, personally, in terms of what it can accomplish for me. And while it could be rightly said I've made things unnecessarily hard for myself approaching life like this, I feel it has been beneficial, and enriching, to force myself to really ask, what is this person saying here? In particular, I wouldn't want GPT to lead to a general lessening of empathy.
> And while it could be rightly said I've made things unnecessarily hard for myself approaching life like this, I feel it has been beneficial, and enriching, to force myself to really ask, what is this person saying here?
I personally don’t believe there’s mutual exclusivity here.
I spent the first half of my life just absorbing. I’ve spent the 2nd half of my life so far undoing the patterns of thought that can result from staying inside one’s head instead of engaging with people.
In my experience, asking the person saying something what they mean is far more effective than asking myself what they mean.
When I ask them, I can form real empathy.
When I only ask myself, I just waste significant energy on all of the ways I imagine I could or should be empathetic.
> In particular, I wouldn't want GPT to lead to a general lessening of empathy.
While I have a lot of concerns about LLMs and the future of literacy, propagation if misinformation, etc, I don’t think ChatGPT is any more risk to empathy than 100 other aspects of modern life.
Facebook, Twitter and Reddit seem far more responsible for the erosion of empathy, and it’s unclear how LLMs would inherently lead to a “general lessening”. I think that ship sailed a decade ago.
Well, here instead of asking you a clarifying question about your response, I might instead ask ChatGPT. Something would be lost, in my mind. Just musing out loud.
I think leaning on ChatGPT to understand others is as problematic as relying only on oneself to understand others, but now you’ve added a layer of your own interpretation depending on how you engage with ChatGPT.
Put another way, if you’re going to sources other than the individual speaking to clarify what they’re saying, the underlying issue is probably not ChatGPT or whatever the next tool is that comes around.
Another form of what you describe is leaning on one’s friends/acquaintances. Plenty of people do this, often with poor results. Reddit’s various relationship forums are a great example. I translate what I thought I heard and ask a 3rd party who wasn’t there what they hear. But by doing so, I remove even more context and make it even less likely to arrive at a useful answer.
I’m sure people will use LLMs for this, but the root issue is deeper, not caused by these tools.
I think that with time, we’ll get better at determining which types of conversations are worthwhile and which aren’t.
If I’m trying to understand a complex multi-faceted technical issue, it’s amazing to be able to drill deeper and deeper into the knowledge contained within the LLM.
If I’m trying to understand the internal states of other people, I have no reason to believe I’ll find good answers in a model that wasn’t trained on that person’s thoughts.
Shortly before GPT and friends burst on the scene, I was looking for a website which would meet me where I am as an engineer - I've written reams of code in various languages.
If I want to try Rust, I don't want to be taught uint8 v uint16 or that you shadow variables. I want to know the interesting parts.
ChatGPT is pretty good at this and the other thing I want: pandas training. You can ask it to generate exercises at any difficulty and also provide test data!
This tool is the biggest mind expander for me since search engines.
I agree, Google has deteriorated so much. It just points to either doc pages and leaves all the work up to you or you get a (often outdated from 2011) pages from Stack overflow. I need to do all the filtering and stitching.
I tend to get better answers from their "automated questions" which are paraphrased versions of my query. So it clearly understands what I'm after.
In order to promote diversity, i would recommend perplexity.ai which offers a similar experience as chatgpt (I'm not affiliated and i have no clue what their tech stack is like) It also offers links back to pages and follow up questions etc. Highly recommended if you need to learn something new and you don't want to bang your head on the keyboard googling or ddg'ing
I'll give an example. I recently needed to learn about k8s, minikube, kubectl et al for a project. I had some vague idea about the tech but nowhere near enough for what i needed to do. Google was useless because it kept taking me to doc pages which is like being lectured but i needed specific information. Perplexity was amazing in helping me with the right bit of information, example code AND links if i do want to read further
I think "What you want to hear" is in the same spirit as an undergraduate attempting to regurgitate half-baked knowledge into an answer that will get him the grade.
Maybe a better rephrasing would be "ChatGPT has been trained to give answers to questions in a clear, confident manner, regardless of the content"
> We trained this model using Reinforcement Learning from Human Feedback (RLHF), using the same methods as InstructGPT, but with slight differences in the data collection setup. We trained an initial model using supervised fine-tuning: human AI trainers provided conversations in which they played both sides—the user and an AI assistant.
So I agree with OP; it's been trained to give answers that sound plausible but not necessarily correct. It's even mentioned in the "Limitations" section at the bottom of the blog post.
> ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers. Fixing this issue is challenging, as: (1) during RL training, there’s currently no source of truth; (2) training the model to be more cautious causes it to decline questions that it can answer correctly; and (3) supervised training misleads the model because the ideal answer depends on what the model knows, rather than what the human demonstrator knows.
It outputs what some idealised version of a person wants to hear, where what is "idealised" has been determined by its training. I've noticed, for example, that it appears to have been trained to want to give responses that seem helpful, and make you trust it. When it's outputting garbage code that doesn't work, it will often say things like "I have tested this and it works correctly", despite that being an impossibility.
Sometimes what I want to hear is "the thing that the typical person who writes about this sort of thing would say about my particular question". If I can figure out whether or not to trust an internet rando who may or may not know what they're talking about, I can figure it out when I'm talking to a simulated internet rando.
Don't forget that GPT was trained on data from all the sources that failed you. So if we all collectively neglect them, and they fall over, we will lose many valuable resources.
I think we need to think about how to keep these valuable sites going, because they are ultimately providing most of the value of the various available language models.
Now whenever I need to use a fancy new package that I never used of, I use LangChain to collect all documents from the package document site, load them into a vectorDB and start asking GPT questions. This method works in 80% of the time.
One pitfall is that with this method I only get what I want. I don't get a deep understanding of that package as it used to be if I carefully read over the documents. Still finding a balancing point in between.
Imagine being able to ask questions and get answers back based upon understanding, and not upon a curriculum or agenda. I have found this method to allow me to consume information much faster while skipping over the often tedious topics.
No,I don't think so. There are times when you ask a question, and the person will condescend, in the positive sense, and speak at a level below you. I prefer to ask when I am teaching where I should start. I'll explain very quickly the basics, if they're familiar, I'll jump the next level. Until I reach the depth of their understanding, I will skip over foundational concepts.
So many comments on this shovelling sh*t on OP...? I agree with the author entirely. In practically every field, the reason it's difficult to learn anything, starting with zero, is because you don't know what questions to ask. You need to find an in. ChatGPT is really good at this - the answers it gives provide an idea of what you don't know you don't know. Obviously, you then need to go verify the answers, but the difference is: now you know what the language you need to use to do so looks like.
PS you would also need to do this if you started with Wikipedia as well.
Finding out what you don't know to start learning is also a part of learning, and it is a bit difficult to trust LLMs like ChatGPT when it had a tendency to spout correct sounding answers that are incorrect, inaccurate etc.
Search engines are much better
On uncovering guides on that, specially from experts and verified sources. It's a bit of work to verify that, but then it is part of learning itself, not sure trying to punt on that to fast track your learning your whatever is going to make a meaningful difference in terms of time
I find this too. With GPT there is no such thing as a stupid question. That is really liberating! You can really double check the simplest and most basic of your assumptions with no fear of judgement. And as mentioned in other comments, the assurance of receiving a good faith answer is not to be underestimated.
im using phind too - phinding it pretty incredible. As someone learning development - it has helped me with so many of the annoyances of learning dev environments, like:
setting up a venv..
environmental variable issues in windows.
*diagnosing a UTF-8 issue in windows.
i get that professionals problems would be harder to answer...however, getting responses without wading through stack exchange entry after entry has really kept me focused, and prevented the often times frustrating recursive spiral which is getting an issue with your issues issue...
It didn't give a wrong answer when asked "Is there a digital to analog converter with an 8V analog range and serial input?", which another poster (mhb) had shown to trip up plain GPT4.
Ahahha here's an interesting thing. I've heard people here beginning to push the idea that learning X isn't important when machines can do it for you, but I was parodying, mocking, and otherwise shitting upon that type of belief, because only a total lunatic or cult member would assert that it's great if you never have to think again. What's indeed disturbing is that you perceive - while I lampoon - that there are many people who may actually hold some partial version of that view.
Sorry. I'm glad I clarified. Almost everything I write should come with a red label above it that says: Gallows Humor. I get taken seriously way more often than I expect to, or should be, around here.
Although it's a little weird that we're on a board where someone could seriously mean what I said and people wouldn't think it was a joke. e.g. the "You let your kids use calculators, don't you?" crowd.
Just now I heard someone unironically say that it's about time the USA government stops the supply of porn to minors via the Internet (presumably that's any nudity to anyone under 21 or maybe 18, another thing I never know when reading USA-originating comments). You just can't tell what kind of wacky opinions are out there and what's a joke. The comment read as though posted by someone who enjoys watching others waste time disagreeing with them, until they said something about social media in the end that's much less far out there and makes me conclude the comment is likely actually genuine.
It's also kinda rude to assume someone is joking when they were actually serious, so yeah one tends to get taken seriously when in doubt...
Here's the aspect of this article that I found the most fascinating, and might explain why I don't get the same level of benefit from ChatGPT as many report.
If I were learning how LEDs work, I would not have wasted any time whatsoever on the search results that the author spent a lot of time on. They were obviously (to me) the wrong articles on the face of it, because they were covering aspects that weren't really what I was looking for (the wrong sort of detail and emphasis).
So I think I would have been off and running pretty much immediately with the web search results rather than spending time on the clear dead ends.
ChatGPT gets me there too, after enough back-and-forth, but it takes longer for me to zero in on what I'm looking for.
I say this not to say that ChatGPT is in any way bad for this. I'm just noticing a difference in how the two of us engage in learning new topics. Perhaps the reality is that for some people, ChatGPT is a godsend, and for others, it's fine... but hardly an improvement for this use use case.
Most people seem to prefer learning by talking and asking questions. At least, that's what I've gathered from Discord servers where 99.9% of people ask questions answered on the first page of the Readme ;)
ChatGPT can be a great supplement for independent research. But when the article mentions a “curious seventh-grader”, I think we should focus on getting them quality human instruction whenever possible instead of just pointing them to ChatGPT.
ChatGPT addresses a scalability problem: not everyone has access to a tutor or can just call up a teacher or mentor to learn and ask questions. But some in the tech industry claim that ChatGPT is as good as or even better than human instruction, which to me seems totally off base.
The biggest problem I see in using LLMs as a teacher-substitute is that LLMs answer the questions you ask, whereas a good teacher tells you what you need to hear. Maybe this is solvable with specialized model tuning, but we need to actually solve it before telling kids that the best way to learn is to talk to the computer.
I totally agree with the author, I find I'm finally able to ask about a range of new concepts, get them explained at my level, and dive deeper as curiosity warrants it.
It's also magical when you summarise the understanding you've reached back to it, and it can confirm or tweak it for you.
In other ways it's also nice to just pay for it and then to be in an advertising free space.
One critique is that when you ask it to compare things it's often too balanced or too positive/enthusiastic ("both are great for different reasons!") when what you want is a more sober analysis. But you can usually do some prompt management to adjust it back to a reasonable range.
This is incredibly, incredibly naive. "Within minutes, I learned" -- absolutely nothing. At most you've learned how the answers to your question would sound like. They are not the answers.
Even if the answers accidentally happen to be correct, that's just the broken clock happens to be correct twice a day. The information value of answers by ChatGPT is zero.
> This is the thing I worry that people are sleeping on. People who think “these language models lie to you all the time” (which they do) and “they will produce buggy code with security holes”—every single complaint about these things is true, and yet, despite all of that, the productivity benefits you get if you lean into them and say OK, how do I work with something that’s completely unreliable, that invents things, that comes up with APIs that don’t exist… how do I use that to enhance my workflow anyway?
That large language models are stochastic parrots is not a bias, it's a fact. You feed it a very large amount of structured data and then put up a question which it'll answer with a series of those which are most likely to follow. There is nothing else.
I know how they work. What's so interesting to me about them is how useful they turn out to be despite the seemingly dumb and simple way they are built.
Yes. Human minds often craft sentences which contain words together which never occurred in the same sentence before, not even close:
> There was a boy called Eustace Clarence Scrubb, and he almost deserved it
> Everything starts somewhere, although many physicists disagree.
> It was a nice day. All the days had been nice. There had been rather more than seven of them so far, and rain hadn't been invented yet
> My father had a face that could stop a clock.
> It is important, when killing a nun, to ensure that you bring an army of sufficient size.
> In the myriadic year of our Lord—the ten thousandth year of the King Undying, the kindly Prince of Death!— Gideon Nav packed her sword, her shoes, and her dirty magazines, and she escaped from the House of the Ninth.
There's just no end to these.
And, of course, science is full of these too, one that jumps to mind is Shinichi Mochizuki's claimed proof of the ABC conjecture which has been proven flawed as it often happens but it was certainly a credible proof despite written in a language no mathematician have ever seen.
Yes but those are as every other thing they emit, indeed , "Let’s be bear or bunny", I heard it called AI hallucination I personally like to call it word salad.
It's easy to cherry pick from the vast human repertoire. Here's a sample from GPT4, where I picked one from a total of two:
> The man on the moon fell off his ladder one Tuesday, a common enough occurrence that nobody really paid it any mind. Truth is, gravity's always been a bit of a show-off, even in the star-dusted emptiness of space.
It's not correct by accident. It's not generating responses randomly, but based on extensive training data. The accuracy of answers is very high, especially on uncontroversial topics like most of the examples in the OP. And the author even described cross-checking the answers for validity and confirming that it was generally correct. (I was also an electrical engineer once upon a time and nothing jumped out at me as wrong, but admittedly it's been quite a while!)
You can't rely in it being 100% correct, but that's very different from it having no informational value at all. When it comes down to it, you can't rely on anything being 100% correct. I recall finding multiple errors in textbooks in the past, and certainly Wikipedia is wrong about all kinds of things; that doesn't make them useless. It just means in situations where it's critical to be correct, you need to double-check. But often that's not necessary, and when it is, it's a lot easier to start with something and then verify it than to not have the tool in the first place.
> The information value of answers by ChatGPT is zero.
Nonsense. The only way this would be true is if ChatGPT's answers were as likely to be correct as random answers. They are much, much more likely to be correct, however, so their information value is greater than zero by definition.
What you're claiming is equivalent to "search engines can find incorrect information, so search engines are worthless for information retrieval". Which is bollocks.
I mean did you read the discussion he had with the language model? Does it clarify anything better than you would get by skimming wikipedia?
I have not found these explanations sufficient, while I have done a bit of chemistry and physics I understand the basics of light emissions, but reading this has the same value as reading the wiki article on LEDs to me.
Sure - when I skim articles, I often don't read all of it and stick to what I'm interested in - I'd argue that even takes less time than chatting with chat-gpt and fact checking it
(Edit: - and you get sources right away with wikipedia!)
Even if the answers accidentally happen to be correct, that's just the broken clock happens to be correct twice a day. The information value of answers by ChatGPT is zero.
Its almost as if ChatGPT is getting to be a form of dialogic learning [1]. While ChatGPT is not yet an authoritative source, it should be possible to get to such a future. The problem is with the ethics of it becoming an authoritative source and how it can possibly unlearn (or accept change or challenges). For now, the example OP provides is without authority where they learnt about a topic taking ChatGPT as a peer.
Fun with caveats I'm finding. Today I asked for an outline of DevSecOps concept based on CISA recommendations. The CISA doc it referenced doesn't actually exist.
Puh, fun, yes. It's nice having a "conversation". But then everything that comes up I have to double check. As soon as the topic is somewhat specific, I can tell just by looking at it that the code won't work. So at the end of the day I have to go back to Google anyways.
Having a conversation allows me to figure out what I’m trying to figure out.
Once I know that, I know what to look for.
To me, the value of the conversation isn’t its perfect accuracy, but the expressiveness and ease of veering in any direction that seems interesting at the moment. The efficiency gains when jumping around within a subject are incredible.
In a similar situation as you, I’ve been having lots of luck using phind.com . They advertise as “the AI search engine built for developers” and its interface is a blend of google and chatgpt. You can also paste in long snippets of code or provide specific URLs you want it to use as a part of its answer. It then generates a few possible solutions with references.
For me it’s consistently been much more helpful than google.
Scary part is I don't know enough about the subject to tell apart truth from falshood because they are stated in exactly the same confident manner. Also most things are true with falsehoods sprinkled in between
I've been using Phind [0] for specifically dev related stuff, and it's been great. I very much like that it gives me all the resources it uses.
On the other hand, I feel like my experience with Phind has been so great because with all dev stuff I ask it it's always been something I could try out right away.
With non-dev stuff that isn't always the case. I'd love an LLM that can do the same thing as Phind, as in providing the sources it's using, while also working for everything.
One solution to his problem with the Wikipedia being hard to learn from is to use the criminally under-known Simple Wikipedia (https://simple.wikipedia.org/wiki/Main_Page). Just replace 'en' in a wikipedia URL with 'simple' and hopefully there will be a nice and simple version of the page. I don't see it advertised very often but I think it should be.
I've found ChatGPT quite useful for both coding and learning a language. Even though I have two friends who are native speakers of the language I'm learning, I don't want to bother them with my questions all the time. Especially since ChatGPT will give great explanations about grammar and why something is said the way it is. Also makes for a great practice partner to use your vocabulary and to correct your mistakes. It has definitely helped a lot already
This is the classic "google is so crap now", then when you dig a bit, the search query is something incredibly broad (in this case "led") or irrelevant, with the author then complaining about the quality of the results
I'm not taking anything away from Chat-GPT for this use case, but I see the above pattern in almost every Google bashing thread on HN.
FYI the autofocus on the API key field was a huge pain due to the Lastpass app (Android); it kept automatically opening the Lastpass app trying to fill it.
Obviously the risk here today is that ChatGPT isn't a trustworthy teacher, but the concept is of course absolutely riveting and obvious to everyone, which is why ChatGPT blew up on the first place. Iron out the kinks and I think it's going to be an incredible asset to humanity (not necessarily this particular LLM).
I was initially using gpt to learn for 5-8 hours every day..
But I was bitten by a series of completely made up facts, and my skepticism has forced me to research nearly every complex response since then to confirm.
This fatigue has really shut down my enthusiasm. Maybe I'll research a topic 3-4 hours a week now.
Hot take: AI knowledge is what Bitcoin is to money. It is not backed by anything and therefore cannot be trusted. Regular knowledge is backed by a degree which has actual value in the end.
im reading the wealth of nations right now, using gpt on the side to help clarify long passages and decode archaic terms. it's been tremendously helpful. some actual use cases:
- summarising entire chapters (summarise chapter 8 of book 1)
- asking to verify my own summaries (in summary, smith writes that ... am i correct?)
- explaining certain passages in ever more detail (explain ...; explain in simpler terms; im still confused, explain in more details;)
---
One very interesting case is when GPT produced an error, and I had it go through the error with me. In the end GPT managed to spot its own error and correct itself. Here's an excerpt, edited for brevity:
* gpt: [...] When the demand is too low, it encourages growth; when the demand is too high, it restrains growth. [...]
* me: "When the demand is too low, it encourages growth; when the demand is too high, it restrains growth." is this correct?
* g: Apologies for the confusion in my previous response. The correct statement should be:
When the demand for labor is too low (slow), it encourages higher wages and population growth; when the demand for labor is too high (fast), it restrains population growth by adjusting wages to find the equilibrium between demand and supply.
(GPT still giving the wrong answer.)
* m: i am still confused. can you explain in more detail
* g: [...] Smith argues that the demand for labor regulates population growth by influencing wages. If wages are too low, fewer children will be born, and the population growth will be slow. If wages are too high, more children will be born, and the population growth will be fast. [...]
(At this point GPT was able to give me the correct response.)
* m: so when the demand for labour is low, wages decrease and population growth slows down. when the demand is high, wages increase and population growth speeds up. correct?
* g: Yes, that's correct. [...]
(I pointed out that this was in contradiction with the earlier response.)
* m: did you not contradict yourself? "When the demand is too low, it encourages growth; when the demand is too high, it restrains growth."
* g: Apologies for the confusion in my previous response. I misunderstood your question. Let me clarify:
When the demand for labor is too low, it restrains growth; when the demand is too high, it encourages growth.
So, when the demand for labor is low, wages decrease, and population growth slows down. Conversely, when the demand for labor is high, wages increase, and population growth speeds up. I hope this clears up any confusion.
(I then inserted the original erroneous response as a prompt, and gpt was able to "remember" that it generated the response earlier.)
* m: "When the demand is too low, it encourages growth; when the demand is too high, it restrains growth." what is wrong with this passage?
* g: I apologize for the confusion earlier. There is an error in the passage you provided. It should be:
This demand for labor operates like the demand for any other commodity, regulating the production and growth of the population. When the demand is too low, it restrains growth; when the demand is too high, it encourages growth. This dynamic determines the rate of population growth in different countries, such as rapid growth in North America, slow and gradual growth in Europe, and a stationary population in China.
---
i should mention that i am using the paid version (GPT-4). i tried the free model, but the quality was significantly worse. you can recorrect it multiple times consecutively and every time it would say that it was wrong the last time, and recorrect itself again and again.
The author describes his learning process in detail and explains what's exciting: he got a basic understanding of the subject within minutes, after being frustrated by Wikipedia, a physics education site, and some other site. Getting through the surface level is a big deal for someone who's new to a subject.
I doubt either of those wishes are going to come true though. Search engines are likely always going to be SEO'ed into uselessness and GPT isn't intentionally telling lies.