Alexis // Designstream

Founder @designstreamxyz | Product Designer

[Alexis // Designstream @theAlexisDS]

@BrettFromDJ 100%

[Alexis // Designstream @theAlexisDS]

@BrettFromDJ That said, they could invest the time in other stuff that need to be redesigned. Maybe I’m old school but I don’t think this way of launching feature after feature every week is good UX long term. I have seen this before and it ends bad. But I guess we are all trying to survive.

[Alexis // Designstream @theAlexisDS]

I thought the same, but I realized some times you update a design directly in framer and it would be good to copy it back to figma. I guess they are thinking for a workflow where you just vibe code and then publish and copy back. But it is true is rare. It is more for stealing lol

[Alexis // Designstream @theAlexisDS]

@BrettFromDJ Do you think that it possible to replace midjourney runway recraft topaz with this one ? I would love to reduce my tools to just one that does it all

[Alexis // Designstream @theAlexisDS]

@BrettFromDJ What was more difficult to achieve? The design or the logic? Did you watch any YouTube videos worth recommending about this topic?

[Alexis // Designstream @theAlexisDS]

Had a fascinating conversation with @cesar_reyesp Reyes about AI.

Not about whether AI will replace designers or developers.

About something much more interesting:

Can you train taste?
I was asking him how he sees AI impacting software engineering and design over the next few years.

His answer went somewhere I wasn't expecting.
He explained that a lot of modern AI systems are not necessarily optimized to generate the most correct answer.
They're often optimized to generate the answer that people are most likely to prefer.
The distinction matters.
Because once you start optimizing for preference, you start asking different questions:

- What happens when millions of people vote on what they like?
- Does quality go up?
- Or do we slowly converge toward the average?

That reminded me of what happened with social media.

A platform starts with one goal.
Then the algorithm discovers what people engage with most.
Over time, the system optimizes for that.
Not necessarily for what's best.
For what's preferred.

The conversation then moved into design.
If creativity is partially subjective, can you actually train a model to have better taste?
Can you teach it what makes a website feel premium?
Can you teach it why one visual direction works for a B2B founder but not for a luxury brand?
Can you teach aesthetic judgment?
Or can you only teach patterns that people tend to vote for?

César suggested something I found fascinating:
Instead of training a model on design outputs, maybe you train it on decisions.
You show it examples.
You show it preferences.
You show it why one solution is better than another for a specific audience.

Not just what was created, but the reasoning behind it.

That feels much closer to how experts actually work.

The more I think about AI, the less I believe the real challenge is generating outputs.
The real challenge might be preserving judgment in a world increasingly optimized for consensus.

Thanks @cesar_reyesp for the conversation.
Left with more questions than answers, which is usually a good sign.

[Alexis // Designstream @theAlexisDS]

https://t.co/5nKwwk2LdN

[Alexis // Designstream @theAlexisDS]

Learning by doing and studying the journeys of people who have built extraordinary companies is my favorite form of education.

Biographies are fascinating because they read like adventure novels, but every lesson, setback, and breakthrough actually happened.

This time, it’s Shoe Dog by @NikeUnleash PhilKnight, the founder of @Nike

One of the biggest reminders from brands like Nike: people don’t fall in love with products, they connect with stories, emotions, and purpose.

Great branding isn’t just about being recognized.
It’s about being remembered.

[Figma @figma]

The first ever Config Makeathon is coming. June 4. $100K in prizes. Powered by @contra

Bonus: pre-register by June 3 for access to Figma’s design agent beta.

[Alexis // Designstream @theAlexisDS]

@BrettFromDJ Congrats looks great ! I will definitely give it a try

[Alexis // Designstream @theAlexisDS]

Short-term future vision: Creatives will be the most important and well-paid again. As UI becomes more about systems that AI can interact with, the old graphic designer role will return to add a creative touch to all types of brands. Branding agencies and digital agencies focused on development (AI-powered software factories) will be the most important. UX will focus only on strategy and product growth, with PM and UX managers as one role.

[Alexis // Designstream @theAlexisDS]

In today’s world, accountability and will, are more valuable than talent.

Stay hungry

[Alexis // Designstream @theAlexisDS]

@RodrigoRomano76 Está la lista ? @RodrigoRomano76

[Alexis // Designstream @theAlexisDS]

I always wonder when AI will be able to see and understand creativity, moving away from logic and “auto layout,” and say “I will put this element here (position absolute) even if it doesn’t follow any logic just because it creates x y z to the overall design.”

[Alexis // Designstream @theAlexisDS]

@BrettFromDJ Great, thanks.

I always wonder when AI will be able to see and understand creativity, moving away from logic and “auto layout,” and say “I will put this element here (position absolute) even if it doesn’t follow any logic just because it creates x y z to the overall design.

[Alexis // Designstream @theAlexisDS]

@FonsMans @figma Congrats man !!

[Alexis // Designstream @theAlexisDS]

@BrettFromDJ trying it now, thanks for sharing