I love science, but I've found that a focus on productivity has softened the general curiosity I started grad school with. To remedy this, my 2023 Resolution is to read a paper a day, focusing on breadth. I'll be tracking my progress and thoughts below 👇

This is exciting; I expect we are going to see a lot more things like this and it will be one of the most important impacts of AI. Congrats to the Future House team. https://t.co/Cxeh8UlWdk

Forbes Hungary listed me among the "25 Hungarians Behind the AI Revolution." I wouldn't be here without the help of my mentors and friends, and I am massively grateful to @FutureHouseSF and @Schmidt_Center @Broadnstitute for supporting me on this journey.

Kosmos gave me a glimpse at the future of science: When I have an evening thought, I pass it to the agent, and by morning I see it validated. Congrats to the team, it was a great pleasure to work together

Today, we’re announcing Kosmos, our newest AI Scientist, available to use now. Users estimate Kosmos does 6 months of work in a single day. One run can read 1,500 papers and write 42,000 lines of code. At least 79% of its findings are reproducible. Kosmos has made 7 discoveries

On November 3rd, join us in Stuttgart: https://t.co/jUxj0PfeGT Or on November 12th, visit us in London: https://t.co/EHl0UeZpGK

Friends in Stuttgart and London — our Biophilia works (w/ Balázs Csizik) are coming to an exhibit near you! See below for more info 👇

Congratulations to our friends at @HHMIJanelia @CamZoology @GoogleAI on the drosophila male CNS connectome!

@E11BIO is excited to unveil PRISM technology for mapping brain wiring with simple light microscopes. Today, brain mapping in humans and other mammals is bottlenecked by accurate neuron tracing. PRISM uses molecular ID codes and AI to help neurons trace themselves. We discovered

TL;DR, I'm looking into ways to accelerate/automate my scientific workflow. Cursor and ChatGPT have been huge thus far, but I'm curious about Claude Code, FH, and Notion. Please share your favorites, or any augmentations on these pipelines! Curious to try any and all.

For the scientific background, I look forward to exploring @FutureHouseSF's paper traversal models, which I've already enjoyed using, but have not tried in a project management setting. I'm open to other ideas here, @NotionHQ has a high activation energy, but seems exciting.

There's room for improvement here as well, but at the very least I get my thoughts back in a structured way, with some next steps and code suggestions. I've found the created ipynb/python code templates and "executive summary" pdfs underwhelming, and cited papers unbalanced.

For project management, I've been using @OpenAI's ChatGPT-5 Whenever I feel "stuck" or "slowed" on a direction, I dump all my ideas into ChatGPT. This either gives me: (1) a new way forward, or (2) a "checkpoint" from which to continue. (h/t @fluorocore for the idea)

In summary, @cursor_ai has been a rocketship for my coding habits and progress, however there still is some room to grow. In terms of what I want to try: 1) Switching out the sidebar agent to improve ipynb compatibility. 2) @AnthropicAI's Claude code, instead of an IDE.

The tab completions and in-text agents of cursor make editing, bug fixing, and even writing from scratch a breeze. The sidebar agent has some room to grow, imo, or I have room to learn. For data analysis, I still like ipynb formats, and the agent often gets lost in there.

First, @cursor_ai. Stunning. I remember the boost in productivity I had two summers ago when I realized ChatGPT can code to some capacity. Then, a year ago, I tried CoPilot and felt it again. @cursor_ai has given me that feeling a third time.

It's our first week of onsite training at @FutureHouseSF, where as postdoc fellows we're exploring how AI agents and tools can accelerate and streamline our research. As part of this journey, I wanted to share my current workflow, and thoughts on how it could be improved 👇

@ForestNeurotech and @ashleevance are still awesome, but the clip can be easily interpreted to mean coma patients are conscious but "locked in." Some may be! And perhaps later in the interview, or in cited pubs, this is clarified. Please send my way if so!

Important point here: the brain processing these signals does not mean that the individual is consciously aware of the speech. I'd be curious to see whether more frontal regions light up, or if memory is being consolidated of the speech being heard.

Fellowship Szn is heating up -- many of the larger cohort postdoc fellowships have due dates October to December. Check out my guide below, and reach out if you have questions!

Gregor is a stellar scientist, and an amazing + compassionate collaborator Go work with him!