Ion Ceban

Tech Lead | Building AI Systems That Scale in Production

[Ion Ceban @IonCeban]

@addyosmani Thanks for the article

[Addy Osmani @addyosmani]

https://t.co/hIe0UX7z6T

[Ion Ceban @IonCeban]

@trikcode Coding is often the easy escape for developers.
The hard part is stepping back, thinking clearly, and deciding the direction.

[Ion Ceban @IonCeban]

@trikcode Hi 👋

Working on applied AI systems, always interested in meeting builders, researchers, and people exploring real-world AI use cases.

[Ion Ceban @IonCeban]

@naval The debate is still theoretical.
Industry needs proof through real products, real workflows, and measurable results.

[Paul Graham @paulg]

When you're deciding what to study in college, don't try to predict what will be valuable in the future, because that's so hard that you'll probably get it wrong. Instead focus on what you personally find most exciting. You can't get that wrong.

[Ion Ceban @IonCeban]

Spun up a clean Ubuntu VM with VMware Fusion today.

Docker
Node
Git
VSCode

Makes it much easier to test things without touching my main environment.

[Ion Ceban @IonCeban]

@kentcdodds I think it’s built through experience.
Complex systems, real users, real failures. That’s where judgment forms.

[Ion Ceban @IonCeban]

@karpathy For me, it’s been simple: I ship more.

I can try new frameworks, test random ideas, and explore things I used to postpone. Everything just moves faster now.

[Ion Ceban @IonCeban]

Building with Codex and Claude Code.

One enforces visible limits.
The other feels almost infinite.

Scarcity changes how you think.

How do you manage AI constraints in your daily work?

[Ion Ceban @IonCeban]

Near-failure training hits different.

[Andrej Karpathy @karpathy]

CLIs are super exciting precisely because they are a "legacy" technology, which means AI agents can natively and easily use them, combine them, interact with them via the entire terminal toolkit.

E.g ask your Claude/Codex agent to install this new Polymarket CLI and ask for any arbitrary dashboards or interfaces or logic. The agents will build it for you. Install the Github CLI too and you can ask them to navigate the repo, see issues, PRs, discussions, even the code itself.

Example: Claude built this terminal dashboard in ~3 minutes, of the highest volume polymarkets and the 24hr change. Or you can make it a web app or whatever you want. Even more powerful when you use it as a module of bigger pipelines.

If you have any kind of product or service think: can agents access and use them?

- are your legacy docs (for humans) at least exportable in markdown?
- have you written Skills for your product?
- can your product/service be usable via CLI? Or MCP?
- ...

It's 2026. Build. For. Agents.

[Ion Ceban @IonCeban]

Good one :P

[Andrew Gazdecki @agazdecki]

Founders when they run out of Claude tokens:

[Ion Ceban @IonCeban]

@netflix Hard words ...

[Ion Ceban @IonCeban]

The goal isn’t fewer tokens.
It’s useful tokens.

Bad context wastes money.
Good context saves systems.

[Ion Ceban @IonCeban]

Working on a social assistant agent these days.
Debugging, testing, breaking, fixing… weirdly fun.

[Andrej Karpathy @karpathy]

I think it must be a very interesting time to be in programming languages and formal methods because LLMs change the whole constraints landscape of software completely. Hints of this can already be seen, e.g. in the rising momentum behind porting C to Rust or the growing interest in upgrading legacy code bases in COBOL or etc. In particular, LLMs are *especially* good at translation compared to de-novo generation because 1) the original code base acts as a kind of highly detailed prompt, and 2) as a reference to write concrete tests with respect to. That said, even Rust is nowhere near optimal for LLMs as a target language. What kind of language is optimal? What concessions (if any) are still carved out for humans? Incredibly interesting new questions and opportunities. It feels likely that we'll end up re-writing large fractions of all software ever written many times over.

[Ion Ceban @IonCeban]

Cardio done.
Quiet mind, ready for the rest of the day

[Ion Ceban @IonCeban]

I notice this every time I do something difficult.

The first minutes feel like a debate.
My mind wants to quit. Very convincing.

Then suddenly it stops arguing
and helps me finish.

Most progress seems to live after that moment.

[Naval @naval]

AI is the great automator, and to automate, it must first imitate.

The imitation fools people into thinking it’s alive.