Pinned Tweet
I'm pleased to present, for free, the first hour of my new course Understanding AI-Assisted Software Development on YouTube!
This course is designed to be agnostic to any one AI tool, and instead provide a mental model and best practices for you and your team. 👇
We programmers have been abstracting ourselves further and further away from the "real code" for decades now.
Memory management used to be a skill until modern languages and garbage collection.
Thanks to vibe coding, all those unnecessary UIs in movies that you couldn't have imagined a real coder taking the time to make...are now totally believable.
2027 will be the era of agentic software factories. Every other aspect of the creation of software will have to adapt. anthropic.com/engineering/re…
First new blog post in awhile. It's a short one, but a response to a lot of conversations I've been having with devs and designers who are unsettled and unhappy as AI changes the industry. tonyalicea.dev/blog/fall-in-…
Tony Alicea retweeted
Claude Code now supports AGENTS.md. We'll fallback to it if CLAUDE.md isn't specified. Hope this helps :)
Tony Alicea retweeted
After co-inventing ChatGPT, I kept asking myself: why have superhuman chat models not led to AGI?
I’ve spent the last 2 years in stealth building a new way to train models (RLCD), and a new type of frontier AI model that we are releasing today: Jev
• 20-200x faster
• 40-400x cheaper (w/ output tokens free)
• Frontier composable intelligence optimized for decisions
AFAICT the shortest path to AI-based economic revolution
Introducing Sloptris! A Tetris clone that satires what AI-generated coding is like.
I've felt a lot like AI-assisted dev shares elements of game experiences. This is a fun little idea I had to express the experience within a game itself. 👇
Gameplay:
- Build blocks manually or "AI-generate" them
- AI tetriminos might mutate after dropping
- Spend valuable time reviewing AI tetriminos to know how they'll mutate
Try it out (beta): tonyalicea.dev/sloptris/
I've been so pleased to see the reaction to the UX.md concept in the design and product community. I'm happy to share an interactive report of my own small experiments with how UX-context impacts product features that AI generates. 👇
Check out the interactive report here: tonyalicea.dev/ux-context/
There's a lot more serious research and training coming from @NNgroup on AI topics - stay tuned!
Every new iteration of software development forgets what UX design actually is, what it accomplishes, why it's important, and how to be good at it. We are at the early part of that again in this new AI phase.
Tony Alicea retweeted
AI makes it so easy to polish rough edges and smooth out your product experience. Yet pretty much every piece of software is falling apart.
Why? The devs building them don't care.
Software built by people detached from the users has never felt worse.
There's a lot of possibility for using AI to build great software. But the simple act of using AI to build software is no guarantee of greatness.
The most popular AI loop seems to be:
1. Human: "I'm bad at writing."
2. <Asks AI to write for them>
3. <AI output is more difficult to read than what the human would have written.>
4. Human posts the AI slop.
5. Human gets worse at writing due to lack of practice.
6. Go to 1.
Tony Alicea retweeted
“Look at my incredible new factory!”
Yo that’s cool, what do you make?
“It’s highly optimised, fully automated, zero tolerance for defects and with a continuous feedback cycle”
Cool cool, so what do you actually make?
“I can interact with it on my phone, laptop, messenger, completely async, and the shared context means it’s always learning how to get better”
Very impressive, but what do you make?
“Every agent has full context, can spawn other agents, review their work, fix defects, and ship continuously.”
yes yes. WHAT DOES IT MAKE?
“Software.”
Oh nice. What software?
“Well right now we’re mostly using it to improve the factory.”
Improve it to make what?
“Anything!”
Such as?
“…a better factory.”
Stop telling me about how many PRs you shipped. Faster implementation doesn’t mean faster thinking.
If anything, it questions how much thought went into what you’re doing.









