ax8080

sure the ones inside the article: "{} simple illustration, naive style, pale colors" the ones in the related / main page: "{} digital art illustration, tetradic moody subdued colors"

Thank you!

Not that it's my merit :)

Can you suggest some cool typography examples? Also, if you could explain a bit about what's wrong with the typography, I'd appreciate it.

I'll contemplate this issue later. I deem this risk to be lower than not discovering a reliable acquisition channel or not being able to attain high margins.

Instinctively it appears as though I'm creating original content. However, all this legal stuff is often counterintuitive.

That would require parsing all of YouTube. I thought I heard that YouTube already did that, but I can't seem to find it now.

Oh, it's okay! In the end, what matters is how much value it gives the user, not how complicated it is inside! thank you!

Nope, Firefox is not planned :( Maybe after Safari, Edge and mobile. Sorry about that.

Yes, I'll probably remove Elon :) Fun fact: twitter banned my ads account, when I tried to create ads with this video.

Thank you! It is very easy for me to make such a service, but then I would lose a lot of users who would install the extension otherwise.

By the way, what browser do you use?

It's not that straightforward. I would say that the most important part of pre-processing is how to break the transcript into parts.

But there are many more things to improve. It's a pipeline: you can add more models, you can train them, you can change prompts, you can post-process the results. But it takes time. More time each step.

Thank you! Yep. Safari and Edge are on my list. It should be easy enough to port there.

It's an interesting idea, but here it's a more pressing issue of copyright infringement.

No, sorry. If I do that, I have to make the plans even more expensive. And users will randomly get free summaries, which will confuse them. "Why should I pay if sometimes I get them for free for no reason?"

Thank you! Yes, I'll definitely do that. That's on the list!

Yes, it's still far from perfect, sorry about that. And it's not about the GPT. I'm sure I can greatly increase the quality on my side.

I have tried many (all the popular) huggingface summarization models, they don't compare to GPT. Basically, they highlight the parts they think are important. But they can't "understand" and generalize.

Frankly, I haven't read the papers about summarization. But I will have to when I'll work on reducing costs.

You can buy the "package" for now :) Subscriptions will be ready by Monday or Tuesday!

I will definitely do that!

and it's funny how sometimes it makes ahhrhrhrhrhhhhh sounds what is the reason behind that?

WOW! What do you use to generate voice? It's SO scary similar to real podcasts. I couldn't find it in a minute in the sources.

Interesting application! It hadn't occurred to me. To my shame, the best I've come up with is to cut ads from influencers videos (make a rewind button).

After installing the extension, go to youtube. There is a "Sign in with Google" button where "related videos" are usually located. If something went wrong, please send a screenshot to me at [email protected]

When I first got to know it, there wasn't much good resources. And now there are 100 times more articles, but I haven't watched them. So I can't recommend anything yet.

Oh yes, I remembered that I saw this - it's good advice: https://help.openai.com/en/articles/6654000-best-practices-f...

Well, I set the limit in OpenAI, yes. But if the service stops working I will also be very upset.

I launched it on Product Hunt last week. I've been tweaking and fixing bugs since then and have now published it here.

Thank you!

I use this this library to do that: https://pypi.org/project/youtube-transcript-api/

I'm really thinking about doing summaries for comments as well.

It's a more interesting task when there are two sides. Even now, my app doesn't work well for debates. It tries to bring the points of view together. I want to work on separation.

Someday, I will make a separate post about it. Although today there seem to be hundreds of people selling "how to work with GPT".

This is what comes to mind immediately: 1. Don't solve more than 1 problem with 1 Prompt. Decompose it into different tasks and make a separate prompt for each one. 2. Use instructions at the beginning. Very short and unambiguous. You have to understand exactly what you want and mean. "Answer me as a philosopher" is an example of an unspecific instruction. 3. If the instructions don't work, show the concept providing an example. Examples are more expensive than instructions because they take up more tokens. 4. The best way to debug propts is when you have a dataset and autotests. I used GPT to evaluate the results. 5. Temperature 0 is fine 99% of the time. (btw I was surprised to find that it does not guarantee a deterministic result; the OpenAI support confirmed that)