> I wasted several hours on occasions where Claude would make changes to completely unrelated parts of the application instead of addressing my actual request.
Every time I read about people using AI I come away with one question. What if they spent hours with a pen and paper and brainstormed about their idea, and then turned it into an actual plan, and then did the plan? At the very least you wouldn't waste hours of your life and instead enjoy using your own powers of thought.
> What if they spent hours with a pen and paper and brainstormed about their idea, and then turned it into an actual plan, and then did the plan? At the very least you wouldn't waste hours of your life and instead enjoy using your own powers of thought.
OP here - I am a bit confused by this response. What are you trying to say or suggest here?
It's not like I didn't have a plan when making changes; I did, and when things went wrong, I tried to debug.
That said, if what you mean by having a plan (which again, I might not be understanding!) is write myself a product spec and then go build the site by learning to code or using a no/low code tool, I think that would have been arguably far less efficient and achieved a less ideal outcome.
In this case, I had Figma designs (from our product designer) that I wanted to implement, but I don't have the programming experience or knowledge of Remix as a framework to have been able to "just do it" on my own in a reasonable amount of time without pairing with Claude.
So while I had some frustrating hours of debugging, I still think overall I achieved an outcome (being able to build a site based on a detailed Figma design by pairing with Claude) that I would never have been able to achieve otherwise to that quality bar in that little amount of time.
I had not thought of visualing my mental debugging process as a decision _tree_ and that LLMs (and talking to other humans) are analogous to a foreign graft. Interesting, thanks!
Never have FOMO when it comes to AI. When it's good enough to be a competitive advantage, everyone will catch up with you in weeks, if not days. All you are doing is learning to deal with the very flaws that have to be fixed for it to be worth anything.
Embrace that you aren't learning anything useful. Everything you are learning will be redundant in a year's time. Advice on how to make AI effective from 1 year ago is gibberish today. Today you've got special keyword like ultrathink or advice on when to compact context that will be gibberish in a year.
Use it, enjoy experimenting and seeing the potential! But no FOMO! There's a point when you need to realize it's not good enough yet, use the few useful bits, put the rest down, and get on with real work again.
If it takes you hours to figure out what's working and what's not, then it isn't good enough. It should just work or it should be obvious when it won't work.
I mean that’s like saying doing normal coding or working on any project yourself isn’t good enough because you put in hours to figure out what works and doesn’t.
That analogy is off, because LLMs aren't a project I'm working on. They are a tool I can use to do that. And my expectation on tools is that they help me and not make things more complicated than they already are.
When LLMs ever reach that point I'll certainly hear about it and gladly use them. In the meantime I let the enthusiasts sort out the problems and glitches first.
I work as a merchant seaman and for our regular day's work everyone basically exclusively uses bowlines, round turn and two half hitches, and clove hitches. We'd use reef knots or single sheet bends for joining ropes.
I think most often people have some vibe coded stuff that kind of does what they want but they don't really understand what it all is and how it works, or any confidence it can be made into something useful, so rather than spending time cleaning up AI code they just use it to grasp the idea and write it themselves. Whether any time is saved by going through this process with the AI seems doubtful to me. Sitting down with pen and paper and thinking through things would probably be more useful.
I think this is a good point because "cheating at the work I have to do, as quickly as possible, well enough to not get fired" is the actual use case for AI for 99% of people.
All the stuff you see in this thread about how kids are going to use AI to bootstrap an education for themselves even better than what their teachers give them (not sure why there's so much hostility towards teachers) is a fantasy.
HN obviously overrepresents kids who were interested in tech things who may do something like that. The vast majority of kids will use AI as a tool to blurt out essays and coursework they don't read, so that they can get back to their addiction to TikTok and Instagram.
As will, of course, everyone using it at work. This is already the case. This is what AI is for. "Do this for me so I can scroll more".
He has two main projects - A & B. Currently A takes up 80% of his time and B takes up 20%. He uses the timers to track time on each project with the goal of transitioning to 80% on B.
If it would have taken you days to learn about the topic well enough to write a bad implementation, how can you have any confidence you can evaluate, let alone "correct", one written by an LLM?
Am I being a completely delusional sceptic or do tools like this make no sense to anyone else?
If you need AI to tell you about a guest and tell you what to ask them, why are you having them on your podcast? You don't know anything about them, and have nothing to ask them - why does the world need you to do this interview? You don't care about your guest at all, so why would we want to listen to you talk to them?
As the guest, why bother responding to questions made like this? Why not have an AI write answers for you, since you are probably equally uninterested in your interlocutor and their questions?
Skip the bs and just publish the transcript of two AIs talking to each other.
> If you need AI to tell you about a guest and tell you what to ask them, why are you having them on your podcast?
Absolutely. The podcasts I like are hosted by people who are always well prepared. For example, Russ Roberts who hosts EconTalk[1] often has guests who have recently published books. Russ reads those books before interviewing them. Amazing dedication.
> If you need AI to tell you about a guest and tell you what to ask them, why are you having them on your podcast?
Podcasts today are little but another appendage of the PR grist mill. Like reality TV, it costs almost nothing to produce yet it can be stuffed with just as many ads. Tools like this help lower the bar even further. Why put out one pod a week when you can churn out 3?
Of the top 100 podcasts today, at least half that are in the "lifestyle" genre where the hosts do nothing besides interview Internet personalities in the wellness, productivity and finance sectors. The pods are 2-3 hours long (more ads can be fit in that way) and I've noticed that the hosts often know zero about the guest and figure it out along the way.
Let's say you recently became interested in X. You don't know much about it. You hear that John Doe is an expert in X.
> You don't care about your guest at all, so why would we want to listen to you talk to them?
It's not that you don't care about your guest. It's that you simply don't know much about X and John Doe.
Is this a reason not to make a podcast at all? I don't think so. Why? Because many listeners might be in your shoes (i.e., not knowing about X and John Doe). In other words... Do you only listen to podcasts when you know everything about the topic and the guest?
> As the guest, why bother responding to questions made like this?
I don't see podcasts as a ring where two egos fight. I wouldn't care about the podcast host's knowledge about my area of expertise at all, as long as they're genuinely interested in it. Isn't this exactly the reason why they invited me? To learn more about it and share it with the world?
I don't think podcast hosts and guests need to be completely "on the same level". PodcastPrepper is able to process dozens of sources from the web in parallel and create a report on the guest in about 3 minutes. If you have 0 prior knowledge about X and John Doe, with PodcastPrepper's report you quickly gain 10x more knowledge about X and John Doe. Enough to be able to make an episode.
Why would you...make a podcast given you know nothing about that person? Or rather, the podcast created by your product is not necessarily "yours." Your product should more accurately be called a somewhat of a NotebookLM by Google, synthesizing data and in the future perhaps creating a podcast from different sources for personal use, not for being production grade enough to publish for others to listen to, as one would expect from a traditional podcast about hosts and guests, like Tim Ferriss, Joe Rogan, etc. Your product is essentially LLM deep research, with text to speech in the future perhaps. This distinction is where most commenters are getting tripped up.
I could easily imagine a scenario where you do know something about a guest - enough to see that they are a compelling speaker, or have an interesting area of expertise that an audience would appreciate - without being fully "read in" on that person.
It's like when you go to a lecture hosted by an institution and they give a little intro about the lecturer beforehand. Most likely some of those little facts about the person's biography weren't known to the institution when they invited the person - they did a little extra research later to prepare the intro. They may not have known what state the person grew up in or where they did their undergrad but they looked it up in case it helps someone make a connection.
I agree as far as my personal listening tastes go, but, I think it’s trying to be a substitute for producer research that bigger shows do for their hosts.
Thanks for the feedback! Exactly! I see PodcastPrepper as augmentation, not automation, of how podcasts are done. Additionally, see my reply to @opto above.
Looks like a great project, and one sorely needed by people like me who find themselves trying to get hold of old books they can't get in their local library and that are too expensive to buy secondhand.
As far as I know Standard gets their raw ebooks from Project Gutenberg which has a vastly greater collection of public domain works. What they're doing is typesetting them for the average reader. But if all you're looking for is just the content, Gutenberg is the place to look for ethically clean copies.
The shadow libraries such as Anna's Archive are a treasure trove of old books, and you're not breaking any imaginary law by downloading old books which are out of copyright.
The internet archive's open library will also link to Standard Ebooks (and Gutenberg and a few others) if a version exists of a book you are looking at e.g.:
> A work that is a mere copy of another work of authorship is not copyrightable. The Office cannot register a work that has been merely copied from another work of authorship without any additional original authorship. See L. Batlin & Son, 536 F.2d at 490 (“one who has slavishly or mechanically copied from others may not claim to be an author”); Bridgeman Art Library, Ltd. v. Corel Corp., 36 F. Supp. 2d 191, 195 (S.D.N.Y. 1999) (“exact photographic copies of public domain works of art would not be copyrightable under United States law because they are not original”).
Certainly! If you add my latest Kirk/Spock slash fanfic to the end of the text, then that is transformative, so the resulting PDF is covered under copyright.
But you wrote "scan". Adding an OCR'ed text layer, or doing manual proofreading and layout ("sweat of the brow") is not sufficiently transformative to have copyright protection.
And we were specifically talking about scans of old books stored in shadow libraries.
I carry 3x5 index cards and a pen with me everywhere I go. When I think of things I need to do, should do, or could do, I write them down. When I get home I see which ones are worth keeping. I put them in a box which has dividers for contexts, like GTD, and a 43 Folders system so that I can schedule stuff for future-me.
That is what you need to get things done every day in a reliable way. Everything else is bs. You don't need it.
More recently people like Casey Reas, who developed the language Processing, Jared Tarbell (https://complexification.net), and Anders Hoff (https://inconvergent.net) are the people I'd look at. Hoff works in Lisp if that's your thing.
Do you know about Peter Struycken, who made many art works, painting, carpets, light shows, and 3D animations with the help of computers? Even his painting of the last decade where made using computer programs to find pleasant looking random patterns for placing coloured squares. See www.pstruycken.nl for more information.
Every time I read about people using AI I come away with one question. What if they spent hours with a pen and paper and brainstormed about their idea, and then turned it into an actual plan, and then did the plan? At the very least you wouldn't waste hours of your life and instead enjoy using your own powers of thought.