This week, I had the opportunity to get a glimpse of what’s coming at the intersection of creative work and generative artificial intelligence at the Upscale Conf in Málaga.
The event was beautifully organized, and I met a ton of interesting people over the course of two days. I’ll surely come back next year.
I took the train back home with mixed feelings: a sense of excitement fueled by gaining clarity about what’s ahead, yet a disconnect with what appears to be an emerging mainstream line of thought I don’t agree with.
1. Extreme delegation. It’s a trope that successful leaders know how to delegate work to others, and indeed finding the balance between multiplying one’s output capacity and controlling the results keeps generating literature.
Enter gen ai: synthetically generated content brings a whole new level of randomness to the mix, and I believe taking this randomness into account is key to making the most out of new technologies available.
I think of it as a cook who doesn’t know what their dish tastes like, yet they’re writing the recipe. One may have to adapt one’s cooking style.
2. Organizations as wrappers. There’s a new model with better capabilities every week. A new startup every day. New tools by the hour. How will we keep up?
As individuals and organizations, we can get inspiration by what the folks at Freepik are doing at the product level: they brought their mission (“help people make better designs”) to the future, and created tools that could live in that future no matter the technology, then switch technologies under the hood as soon as they’re available.
Because they’re always ready, plugging a new model to their upscaling feature means changing a couple of lines of code.
On top of the constant upskilling we’ve grown used to, rethinking our workflows as ‘wrappers’, and being ready to plug new available tech at any time, will help us succeed.
3. Findability, again. I started getting the help of AI for coding about six months ago. Since then, the tooling landscape has evolved so much I lost track of IDEs, copilots and full-blown platforms that “I should have tried”.
That thing you built last month? The way you did it? It probably existed. Or there was likely a 10x better way to build it.
Who’s gonna help us find the right solution for our needs? There’s your free startup idea.
1. Don’t take the fun off of me. I’ve said it before: there seems to be an agreement that AI can help us think, and I disagree (in Spanish, sorry).
Take this quote from Lenny’s How AI will impact product management article:
Instead, I believe that AI will have the most profound impact on the high-level (and historically most valued) skills of product management: developing a strategy, crafting a vision, identifying new opportunities, and setting goals.
Again, I disagree.
During the conference, some seemed to build on this idea of AIs thinking for us: from coming up with fully developed startup ideas to synthesizing the result of anthropologic research field work.
I won’t argue about whether an LLM can better understand people problems than I do (I believe it can’t), but I’ll say this: I don’t want to miss the fun.
To me, hearing leaders and entrepreneurs (who otherwise wouldn’t lower their Founder rank to follow someone else’s mission) be excited about blindly following a GPT idea vomit is sad.
2. The power of team work. If you have free assistants who not only help you achieve more faster, but complement your work with things you can’t do… then you can deliver work by yourself.
At least that’s the idea these days.
I had heard of the one-person unicorn wet dreams (who hasn’t), but in Málaga I was told something different: that by getting rid of intermediaries, the work would be better.
The idea here is that a genius’ seminal concept is corrupted by team work, and so that by removing collaboration out of creative work, quality will increase.
Well, I have yet to meet a genius whose idea cannot be enriched by others’, or a leader who can direct the work of all disciplines involved in any relevant project. Don’t get me wrong, I am an optimist believer that a lot of value can be produced by a single person. But we are talking about creative quality. What kind of teams have y’all been collaborating with? In my world, work improved so much when the right talent was brought into the mix for true collaboration! And, then again, where’s the fun of going solo?
3. AI generated graphics are ugly. And they all look the same. And they are often sexist. Sorry, it had to be said.
Do you remember how, when you were a kid, and you painted with all the crayons at hand, you ended up with a brown color that looked like shit? It seems to me that AI generated art suffers from the same effect.
I am not going to get into the details of why (there are countless articles about it), but it’s so hard to find synthetic images that bring a truly original perspective.
Multi-modal generation tools, where an image — or an image plus text — can be used as a prompt, and iterative processes, where you can refine previously generated graphics, seem to make it easier to get better results, but so far AI art has brought to our screens what large-format digital printing did to high-street store signage. We’re still mourning neon.
And please, enough with the sexy young ladies with robots.