Danilo Gasques, Ph.D.

Looking for the next problem worth solving. Previous: @southpkcommons Founder Fellow @medivis_ai @IntuitiveSurg. HCI/CS Ph.D. @DesignLabUCSD

[Danilo Gasques, Ph.D. @danilogasques]

Initially, LLMs made me a 5x engineer because I could instantly tap into the knowledge of every engineer in the world to stress‑test my designs. Then they made me a 100x engineer because they started writing full applications in a few hours

But there’s a shift: AI is no longer just making us more intellectually competitive, it is starting to replace a lot of the learning and work we used to struggle through ourselves.

Faster end-to-end systems: yes - but what about the human in the loop?

[alphaXiv @askalphaxiv]

Introducing autoresearch for arXiv papers

Change 'arxiv' to 'autoarxiv' in any paper URL

An agent deploys to resolve setup issues on the codebase, run a minimal reproduction, and estimate full replication cost. Read more below

[Danilo Gasques, Ph.D. @danilogasques]

@AgrawalArian 🔥 🔥 🔥

[Danilo Gasques, Ph.D. @danilogasques]

@dina_yrl This is so interesting - in a way, full circle back to where it started (https://t.co/qnMP8rfvVo)

[Danilo Gasques, Ph.D. @danilogasques]

I’ve been offline and traveling for weeks, which pushed me to obsess over running a local LLM coding agent. I wrote it by hand (~680 lines of code).

The hardest part has been working around a tiny context window, and while @lmstudio has been neat, it still has some quirks I'm trying to tame.

[Danilo Gasques, Ph.D. @danilogasques]

@cjzafir This is so impressive. Meanwhile my self-boostrapped coding agent gets stuck procrastinating when the task is not fully articulated:

[Danilo Gasques, Ph.D. @danilogasques]

@lmstudio This is wholesome. Does it have function call support? We are one step away from a token-free LLM that works better than @Apple's Siri :D

[Fei-Fei Li @drfeifei]

https://t.co/Kt50ttQRMJ

[Danilo Gasques, Ph.D. @danilogasques]

@xiaoyiz_uw I love that I somehow get all @claudeai news here before I see the official @claudeai tweet / X post :)

[Danilo Gasques, Ph.D. @danilogasques]

"Bad speed is impatience that pretends to be ambition. It optimizes for novelty over depth and doesn’t let you sit long enough inside the unglamorous stretch where the real edge is formed.

**Because the truth is that the frontier rarely yields itself in the first six months.**"

[Aditya Agarwal @adityaag]

https://t.co/AoyPtuI9jx

[Danilo Gasques, Ph.D. @danilogasques]

@OmidMogasemi 😭

[Danilo Gasques, Ph.D. @danilogasques]

@jakeottiger Super interested in what you are cooking there @jakeottiger !

[Danilo Gasques, Ph.D. @danilogasques]

In most navigation systems, surgeons need a rep at their side just to drive the software.

At @medivis_ai we flipped that: surgeons focus on the surgery, and the system quietly guides them with a structured, step‑by‑step workflow.

That shift in UX was my first big contribution there.

[Medivis @medivis_ai]

We're proud to announce a significant step forward in neurosurgical care. Dr. @RandyDAmico_MD has performed the first catheter placement navigated with Medivis' Cranial Navigation platform. It's a first that matters — intracerebral hemorrhage is one of the most time-sensitive and high-stakes conditions in neurosurgery, and precision is everything.

Post-op imaging confirmed accurate IRRAflow placement within the target hematoma, facilitating efficient clot evacuation.

Medivis' AR-guided platform delivers immersive, real-time anatomical visualization directly to the bedside and OR — giving surgeons the precision and confidence to navigate complex anatomy with clarity.

Dr. D'Amico's work exemplifies what's possible at the intersection of cutting-edge technology and surgical expertise.

Precision. Confidence. Better outcomes.

This is what the future of neurosurgery looks like. Congratulations to Dr. D’Amico and the entire team!

[Danilo Gasques, Ph.D. @danilogasques]

Help! After @Google deployed Gemini to my Google Home devices, they burn a forest every time I ask them a question and yet, answer incorrectly half of the times

Huge latency to start answering; repeating the last answer for a completely new question; giving the wrong information

[Danilo Gasques, Ph.D. @danilogasques]

@AgrawalArian I love this provocation.

If the future of computing is "upload notes -> generate deck," I'm walking into the ocean.

[Danilo Gasques, Ph.D. @danilogasques]

All I want with @figma was to have a consistent design system a non-developer can chime in, tweak, and make it consistent across views. But they made it so hard to collaborate with any agent.

@claudeai #Design nailed the pain point.

Thanks to @vitorialima_42 for pointing this out!

#BuildInPublic

[Danilo Gasques, Ph.D. @danilogasques]

A couple of months ago I spend a painstaking amount of time trying to tweak a website design using @claudeai and @figma - it never worked. MCP was read only or required expensive licenses.

Now, in 5 minutes @claudeai Design recreated all the views directly from source code. 1/2

[Danilo Gasques, Ph.D. @danilogasques]

@cdossman @snowmaker Who is we? @cdossman Curious about this solution

[Danilo Gasques, Ph.D. @danilogasques]

@anandx Pretty slick! Great choice of open tabs for the video: really tells the story of someone navigating thousands of ideas and designs without hunting one-by-one for that tab they opened 100 tabs ago.

[NASA @NASA]

Liftoff.

The Artemis II mission launched from @NASAKennedy at 6:35pm ET (2235 UTC), propelling four astronauts on a journey around the Moon.

Artemis II will pave the way for future Moon landings, as well as the next giant leap — astronauts on Mars.