51–60 of 222 posts
TBH, at this point, if there is a problem that hasn't been solved, its because some people tried and failed and found that the problem is not solvable currently. It may be more fruitful to start with something promisingly popular, and improve upon it (e.g. tiktok)
This essay is very short. Has he given up on writing? Maybe too much Twitter has melted his brian.
I really liked the length of this “essay”.
If he were to make it longer some examples might be interesting to read about and drive home the message, he no doubt has examples in his mind, and his readers would no doubt be curious as to what they are
Question. How does one feed content to ChatGPT and get such output?
not an expert, but i think usually one of two ways - either use openai’s finetune api (the right way) or embed all the docs, retrieve them using embedding of the questions, and add just enough of them to the prompt to generate a response (gpt-index or langchain help automate some of this). downside of former is lack of sources, downside of latter is context constraint
{"prompt": "", "completion": ""}
So I'm not sure how you would feed it PG's essays in this mechanism
i am genuinely surprised that pg is writing such a short response to a question that seems right down his alley , seemingly very core to his subject matter. how has he not written an authoritative “how to get ideas” post until now?
When I was in my 20s, I could code for days on straight. I had so much energy and the perception of the bottomless capital of time. Now that I am forty, the cost-benefit analysis is getting tougher. Can I have my big break and retire early? Sure, but the chances of that happening are not that great. Many serial entrepreneurs fail multiple times with nothing to show for it. All I think about now is - when I am on my d…
* Overture, Picasa and many more OGs.
"ChatGPT, gimme 10 startup ideas" TBH, at this point, if there is a problem that hasn't been solved, its because some people tried and failed and found that the problem is not solvable currently. It may be more fruitful to start with something promisingly popular, and improve upon it (e.g. tiktok)
Real progress in any field or endeavor comes down to about 5% new ideas and creativity, 95% hard work putting those new ideas into practice or building something functional out of them. That 5% is pretty critical however, as without it you're just continuing to redo what's already been done.
This is very true. Ideas are pretty cheap to come by. Implementing, building, teasing them out into a practical scenario is hard.
It starts with finding the right problem/need and before jumping into to solutioning, the smart person would validate and pitch the problem to as many relevant target customers as possible. Once the problem is validated and tangible root cause is flushed out then it's time to explore the solution and pitching it to the same customer research group for validation before going into building mode.
Finding ideas is not hard, but the process is gruelling, frustrating and potentially rewarding. Good luck!
Short and concise, this combines two of my most favorite mindshares.
1. "Hackers and Painters" – I stumbled across in late 2004 at the Technical University of Vienna's (excellent) bookstore. Reading it changed my life.
2. "Ideas are the Enemy of Observations (2012)" – Just Three months ago yamrzou shared this gem of a text. For me, it has become the most meaningful post ever shared on HN.
"How to get new ideas" seems to fit snuggly right in the middle :)