Your reflection, Freddy, is quite accurate and aligns with many of my beliefs about what AI means. It seems the trend is to reduce the number of employees in any company because if we use AI, boom, increased productivity. That's not the whole story. What is undoubtedly happening, and what we should be aware of and understand the implications for our jobs, is getting swept up in this wave of agents, of vibe coding. As you rightly point out, the result the model gives us is random; good result... bad result, we don't know. We have to dive in and see what these companies' data centers produce.
It's instructive that the the "probably most successful lovable-built app so far" according to Lovable's own co-founders is $3M USD in sales for a company **selling a previously planned offered** to their **existing customer base**.
Big success indeed. But clearly a case of 'cost remover' (faster launch for planned upsell) rather than 'revenue adder'.
I think I had my vibe coding addiction phase much earlier, in late 2023. My CEO is going through it now, and it's amazing because he's also producing a lot more projects in a shorter amount of time.
For my part, the takeaway is that I try to reduce my involvement in coding. I dedicate my main focus to other tasks, like planning, auditing, creative work, and learning the most difficult things.
I forbid myself from doing anything after 10:30 pm. I can leave an agent working with the `/goal` at most, but working after a good night's sleep is the best.
Your reflection, Freddy, is quite accurate and aligns with many of my beliefs about what AI means. It seems the trend is to reduce the number of employees in any company because if we use AI, boom, increased productivity. That's not the whole story. What is undoubtedly happening, and what we should be aware of and understand the implications for our jobs, is getting swept up in this wave of agents, of vibe coding. As you rightly point out, the result the model gives us is random; good result... bad result, we don't know. We have to dive in and see what these companies' data centers produce.
It's instructive that the the "probably most successful lovable-built app so far" according to Lovable's own co-founders is $3M USD in sales for a company **selling a previously planned offered** to their **existing customer base**.
Big success indeed. But clearly a case of 'cost remover' (faster launch for planned upsell) rather than 'revenue adder'.
reference: https://www.linkedin.com/posts/antonosika_lovable-built-app-just-made-3m-in-48h-probably-activity-7338217572556795905-aD3O/
another angle to this I think too is: as it's usually used, LLMs are a 'negative remover' instead of a 'positive adder'.
It's good at removing frictions; but not particularly good at adding positives.
Related (but not the same): it's good at cutting costs, but not particularly good at adding revenue.
When we spent a lot of time vibecoding new projects; perhaps we're doing a good job cutting costs/removing friction -- all good things.
But often the bigger job of the CEO is to add revenue.
(thinking this out loud, so this is still rough)
I think I had my vibe coding addiction phase much earlier, in late 2023. My CEO is going through it now, and it's amazing because he's also producing a lot more projects in a shorter amount of time.
For my part, the takeaway is that I try to reduce my involvement in coding. I dedicate my main focus to other tasks, like planning, auditing, creative work, and learning the most difficult things.
I forbid myself from doing anything after 10:30 pm. I can leave an agent working with the `/goal` at most, but working after a good night's sleep is the best.
PS: I loved the Fable 5 course!