Do you use AI for brainstorming ideas? Since yesterday I stopped.
Before AI I used to call my brother or meet with a friend and discuss over a few hours an idea that is blurry in my head. I would start pitching and they would ask questions, because they wouldn't understand it well. I would think about answers and that would help me clarify the idea in my head.
By the end of the meeting I would have a clear vision of a product that this initial blurry idea was meant to be.
Kind of like filling out @ycombinator form just to ask myself some hard questions to help evaluate the startup idea.
With AI though, the blurry idea doesn't really get enough time and thought to get naturally clarified in my head.
The agent immediately spits out the final version of the product he understood from my initial unclear prompt.
Even giving an agent instructions to ask questions before making any conclusions feels fake. I see that answering these question doesn't really lead me anywhere.
It really feels like some processes deserve to take time and not be accelerated by AI.
The only helpful part of idea brainstorming with AI is asking it to find potential competitors and summarise their products.
One more rejection in the bag.
TUM ai has launched an E-Lab initiative in 2022 helping early stage founders based in Munich to build their startups. It is a 12 weeks equity free accelerator program. Since then every year it has been only growing.
If you are in Munich and thinking of becoming a founder I suggest applying to the next batch. Being equity free makes it worth applying for anybody at the very early stage.
@StephNass It is on point. But somehow I have a feeling that most of the things in life are following this pattern. Whenever there are two parties in any relationship this chemistry shows up.
@lukOlejnik Investigation will show. If it was social hacking or someone who had an employee access for example - then it is a crime.
But if it is someone discovering the loopholes in the firmware using AI - then we are experiencing negative sides of rapid uncontrolled growth of AI.
@SantoshYadavDev Great advice. In fact I would say that staying at your job and position is by default degrading after the first year or so.
Once you get comfortable with what you do, it is time to prepare for the next move be it inside or outside of your org.
But that’s not vibe coding you are describing. That’s programming.
A true vibe coder should be okay with 5 different variables representing the same value created at different times by AI through a crappy prompt.
Why?
Because he is not even aware of it. The moment he opens up a black box, he discovers not only that but all sorts of inefficiencies and redundancies and he probably transforms into a real software engineer.
Track status of your md files as colors and icons in VS Code hierarchy. Make your agents workflow easier.
And it's free: https://t.co/TWYPQ6drXu
Tell your friend! See ya
#gamedev#VSCode#DevTools
Don’t you think it means a different way of showing code changes should be presented? Summary? Too shallow, code? Too long and hard to understand at yhis point. There must be something in between. A little technical and a little abstract.
I am trying to build that output for frontend code at least.
@claudeai rejected my startup application for free credits.
Although the reason is not correct, since Abbox does fall into the first category.
I think they stopped the program due to unexpectedly high volume of applications and people applying as a startup but using it for personal needs, so reapplying doesn’t really make sense.
I don’t know if it helps non risk takers to actually take risks.
What I noticed is people who are inclined to not taking risks over a long period of time end up designing their life in a way that taking a risk is nearly impossible. And they justify not taking a risk as the best option they have which happens to be right.
On the other side people who love taking risks somehow end up living a life in which taking the next risky step is somehow the most rational thing to do.
The point here is that if amazon doesn’t let AI agents shop. Then somebody else can create a tiny store and let AI shop there and by doing that learn. The most important asset in today’s AI world. Learn how and what consumers and their AI agents want to buy on amazon.
Nobody has to replace amazon, but creating a highly specialized app that helps AI agents do basic shopping is quite important.
I think amazon will eventually allow it. It is just not ready yet, as its search is complicated and designed for human eyes.
So I recently integrated AI into my startup abbox.
And I thought, perhaps there are some available credits that can cover my AI prompting expenses.
I came across these 3 big players
@Google with $300 (€257.47) cloud credits
@Microsoft with $200 cloud credits
@AWS with $1000 cloud credits
Although I got all three, it is obvious, I ended up using AWS.
I wonder if the math behind giving away free credits works for these companies and if it does why such a huge difference in the amounts?
@StephNass I think the right email will not be read until the end by a human. There is definitely some AI summary tool used by Johny VC to not miss opportunities that are hiding between proses but also to not waste his time reading through thousands of such emails.
It is a genius way of bringing potential customers in at an early stage. I think it is quite cheap, instead of running ads or paying for other ways of marketing.
People will naturally lean to try things that are free and while they are chatting, it is in your hands how to subtly direct the conversation towards conversion.
Even if a small percent of free chatters end up clicking that build button, you got yourself a good deal.
Very well thought approach to customer acquisition. More companies should follow your lead.
@a_shimanski The goal is noble the motivation is questionable. You should first find out if locals in these countries do “small talks”, especially with non-locals.
@Yuchenj_UW Does that mean most of the apps that you need to be controlled on your phone do not have their corresponding web-apps?
If they do have them, then perhaps there are other ways to achieve the same goals.