Arturo Rios (@arturoriosme) 's Twitter Profile
Arturo Rios

@arturoriosme

🇲🇽 Banjo enthusiast 🪕. Product Designerd at JOYA. Writer. Chill as Folk. Human, all too human. Everything we design is designing us in return.

ID: 2907333293

linkhttp://www.arturorios.me calendar_today06-12-2014 05:32:50

1,1K Tweet

317 Followers

1,1K Following

Arturo Rios (@arturoriosme) 's Twitter Profile Photo

If you're a designer and want to start building your own software, here's my list of what I will be focusing on mastering for 2026: 1. Claude code 2. Claude Agent SDK for building agents 2. Ralph Wiggum for iterative work 3. Convex for databases 4. Compounding engineering

Arturo Rios (@arturoriosme) 's Twitter Profile Photo

We need a Ralph Wiggum not for building while we sleep, but for finding what is actually worth building. Everyone focusing on shaping and building but not framer. Makes sense, as there is not instant gratification in realizing our ideas are crap. But there is instant

Arturo Rios (@arturoriosme) 's Twitter Profile Photo

Huge fan of the askUserQuestion UI on Claude Cowork, but honestly, I think the background grid is unnecessary. I can see why they added it (to allow users to know they're on cowork, not on chat or code) but it is distracting.

Huge fan of the askUserQuestion UI on Claude Cowork, but honestly, I think the background grid is unnecessary. 

I can see why they added it (to allow users to know they're on cowork, not on chat or code) but it is distracting.
Arturo Rios (@arturoriosme) 's Twitter Profile Photo

Granola is overrated, tbh. On the other hand, Krisp is extremely goated. Not only does Krisp improve my sound quality, but also has higher transcription quality. Comparison: Krisp is the higher-quality transcription, it has clearer speaker separation, more faithful

Arturo Rios (@arturoriosme) 's Twitter Profile Photo

OpenAI really thinks people will happily download and actively use 3 Mac apps now: ChatGPT, Atlas, and now Codex. This is why Anthropic is always 2 steps ahead. They meet their users where they are and work on frictionless experiences.

Arturo Rios (@arturoriosme) 's Twitter Profile Photo

Claude code v2.1.68 proactively decided to create multiple agents in multiple worktrees to parallelize work. Pretty awesome. Before I would sit alone and wait for things to be complete. Now I have company: the main agent is also sitting and waiting with me.

Claude code v2.1.68 proactively decided to create multiple agents in multiple worktrees to parallelize work. Pretty awesome. 

Before I would sit alone and wait for things to be complete. Now I have company: the main agent is also sitting and waiting with me.
Karri Saarinen (@karrisaarinen) 's Twitter Profile Photo

I think we have lost some sense of judgment and moderation when it comes to product building currently. The moment you turn something into a universally celebrated metric, whether that is token burn, prototype count, or percentage of agent-written code, you start losing sight of

claire vo 🖤 (@clairevo) 's Twitter Profile Photo

What's happening: 1 PM managing 20 engineers worth of output. What it means: if your PMs were mediocre before, they're now mediocre at 5x scale.

Arturo Rios (@arturoriosme) 's Twitter Profile Photo

Claude Code did not get nerfed, but boy V2.1.104 is really slow. Getting to this message almost every time. cc Boris Cherny in case this is useful. Experiencing this with both an Enterprise and my Max account

Claude Code did not get nerfed, but boy V2.1.104 is really slow. Getting to this message almost every time. 

cc <a href="/bcherny/">Boris Cherny</a> in case this is useful. Experiencing this with both an Enterprise and my Max account