AI & Tech Explorer · 15+ years building meaningful software · Digital Artisan & Musician · Founder & Engineer at @zenjoy · Building zentty.org
🇪🇺
- Tweets 19,153
- Following 424
- Followers 2,787
- Likes 21,351
Pinned Tweet
The Founder-Parent Paradox:
You're building a business to provide a better future for your family, but the act of building it takes you away from your family in the present.
The ultimate balancing act.
19
2
136
17,515
Both are child's play compared to explaining to the client why their "simple" change request will take 6 weeks and a full database migration.
5
9
282
27,784
This is how I see it:
From The Browser Company's side, this is about resources and risk. They need a massive war chest, someone with big pockets, to compete with Chrome/Gemini and Perplexity. This provides that. Plus an elegant off-ramp for investors after the pivot from Arc to Dia likely alienated a lot of their core fanbase.
From Atlassian's side, it's a huge bet on the future of work. They acquire a world-class ux team and a head start in the agentic browser race. Imagine an AI agent that integrates deeply and fluently with Jira and Confluence.
It's a strategic play to own the interface where work gets done: the browser.
6
5
244
81,131
First, it solves our hardest tests. Next, it will solve our hardest unsolved problems.
18
12
215
214,224
The last few years of AI drama are going to fuel a decade of Hollywood blockbusters. The egos, the money, the backstabbing, the mistakes, the victories... it has everything.
3
8
162
24,637
Isn’t this similar to the early railroad era? It was a brutal investment for many initial shareholders, but the infrastructure they funded unlocked a century of value for the entire economy.
Society can win even if the first investors lose.
16
6
143
7,596
"We are not building animals. We are summoning ghosts."
This framing is going to stick with me.
3
5
132
14,346
You should find those users, hire them and put them in the stress-testing team. Win-win!
3
117
11,882
Just breathed new life into a 9+ year-old MacBook with @dhh’s Omarchy.
After a bit of tinkering to get wifi, audio, and the touchbar reliably working (can’t live without that escape key), it’s running smoother than ever.
Feels good!
10
93
10,978
Just discovered @dani_avila7's Claude Code Templates: it includes a powerful CLI and web UI for configuring Claude Code projects and monitoring token usage or insight on each tool invocation in real time!
Really nice! 👉 github.com/davila7/claude-co…
3
13
93
4,350
The best model doesn't automatically win. The best *product* does. And Google’s apps are still lagging in terms of user experience compared to ChatGPT or Claude or Grok.
9
1
78
7,684
“Elixir is the best programming language for building AI agents... because agents they are real time collaborative.”
Great interview with @josevalim by @evilmartians: piped.video/ihn2fuc_ueQ?si=LXu9…
20
74
11,133
A bash-mode, so you can execute something manually and the agent immediately has all the output as context.
4
73
3,342
Famous last words of so many acquired startups: "We will operate independently."
I'll check back in a year.
1
1
72
10,974
My wishes 🙏
Fix: the ability to resume previous sessions (either closed accidentally or deliberately).
Productivity unlock: sub-agents like in Claude Code.
Keep up the good work!
1
65
6,643
This is the painful truth for so many long-time Apple fans.
We bought into the ecosystem for the “it just works” magic. But these days, the gap between what's possible and what Siri can do is really becoming impossible to ignore.
2
59
2,738
Hold out if you can! The M5 is rumored to get the A19 Pro's dedicated matmul acceleration, making it a beast for running local LLMs.
5
61
5,145
This is the brutal math of venture-backed, platform-dependent businesses that is often hidden behind impressive revenue numbers.
A good breakdown of why “owning your distribution” means everything.
1
53
20,126
I was thinking the same thing.
"Active user" = simply accidentally tapping the Threads button on Instagram.
46
1,769
They won't call it "choosing colors."
It will be "a strategic pivot towards user-centric personalization to enhance platform delight and drive engagement".
41
1,725
So far, it's not disappointing. But please... never again invert a progress bar silently. 😂
Nothing wakes you up faster than seeing a maxed-out weekly usage bar ten minutes after trying a new model update.
Turns out the UI team decided to invert the usage graph without warning.
Please... never invert a progress bar silently.
2
1
44
48,488
Great list. I'd add "human craftsmanship" to it. Hand-made goods, unique art, or anything authentic and difficult for AI to replicate.
My question is how to invest for the chaotic 10-20 years it might take to get to a post-AGI world. The transition will be a rollercoaster, no?
2
42
12,096
This is in our human behavior: people want to associate with success.
When you're struggling, helping is a risk. When you're succeeding, giving opinions is an easy way to feel part of your story.
2
8
42
1,811
This is the core debate in AI right now.
The big question is whether LLMs are a dead end for AGI, or a foundational component on which future architectures will be built.
LeCun has clearly made his bet.
10
1
40
18,542
The “deep systemic failure” is that a child is now our generation's most expensive luxury good.
1
39
2,089
For me, it made web dev feel fun again.
Instant feedback of your changes makes development truly enjoyable.
1
41
5,450
I have been using @diabrowser for a couple of weeks now, and it's amazing how quickly you take for granted how its real-time context from open tabs supercharges your productivity and sparks creativity.
Here are my top 5 Dia skills I've been using daily 👇
5
2
38
6,283
Isn't this the circle of life for dev tools?
Convenience creates adoption, adoption enables lock-in, lock-in inspires the next open-source alternative.
2
2
38
60,316
Output is only half the battle. The other half is inventing an input method that doesn't make you look ridiculous in public.
21
32
13,243
Prescribing antidepressant shouldn’t be taken lightheartedly.
But there are also people that genetically have much lower base serotonin levels than the average person. For those, an SSRI really means the difference between living life to the fullest or having no life at all.
13
34
83,667
I have one on iOS that's been really helpful:
1/ take the url from Safari or YouTube app
2/ combine this with a prompt to extract any interesting frameworks or novel ideas from the link
3/ put the prompt in my clipboard
4/ open Gemini
... and I can just paste and submit! ✨
2
33
35,924
This feels like the "photography vs. painting" debate of the 19th century.
Early photos were seen as a "cheat" for not requiring the painstaking work of a painter.
AI world-building might be a new art form entirely. The wonder may not come from the lore itself, but from the artful way a human directs the AI to create it?
2
3
29
10,304
Great explanation.
I always think of it like an open book exam:
A standard LLM is taking a traditional test, relying only on what it memorized during training.
RAG gives the LLM an open book, allowing it to look up the correct answers in the textbook before writing its reply.
2
32
3,483