John Giorgi

Enjoyed talking to @SteveLohr for this piece on AI-powered coding assistants and their impact on SWE jobs. To add some color to the point about junior devs: it's a brave new world, but there will always be work for smart, driven & caring people (on that note, Abridge is hiring!)

you can outsource your thinking but you cannot outsource your understanding

“all 4202 tests are passing why is there a bug?”

@sloppenheimer Hey! We wrote a paper using a very similar eval: arxiv.org/abs/2306.11167

Gemma 4 31B is now officially the smallest model that has ever passed my NYT Connections eval suite. It notably did not get the last "category" and ended up calling it the "Purple Group". It took ~250k tokens of reasoning and 7 minutes, down from o1-preview taking 35 minutes.

I'm not very happy with the code quality and I think agents bloat abstractions, have poor code aesthetics, are very prone to copy pasting code blocks and it's a mess, but at this point I stopped fighting it too hard and just moved on. The agents do not listen to my instructions in the AGENTS.md files. E.g. just as one example, no matter how many times I say something like: "Every line of code should do exactly one thing and use intermediate variables as a form of documentation" They will still "multitask" and create complex constructs where one line of code calls 2 functions and then indexes an array with the result. I think in principle I could use hooks or slash commands to clean this up but at some point just a shrug is easier. Yes I think LLM as a judge for soft rewards is in principle and long term slightly problematic (due to goodharting concerns), but in practice and for now I don't think we've picked the low hanging fruit yet here.

The head of @AnthropicAI's biology and life sciences business recently spoke about the company’s approach to building and partnering in healthcare. “In ambient AI, for example, Abridge is already a partner,” said Eric Kauderer-Abrams, Head of Biology and Life Sciences at Anthropic. “Our perspective is that we develop products where we see a gap. If there’s a great product already serving certain use cases, such as Abridge, we don’t need to reinvent the wheel.”

claude, make strawberries sweet again. bring back the warmth of the summer sun when the days stretched on forever and all we had was each other. do not make mistakes.

@xianjun_agi @johnschulman2 @AbridgeHQ is hiring!

I was laid off by Meta today. As a Research Scientist, my work was just cited by the legendary @johnschulman2 and Nicholas Carlini yesterday. I’m actively looking for new opportunities — please reach out if you have any openings!

Modern AI teams need hyperscalers & neoclouds, but legacy tools like SLURM can't keep up. @AbridgeHQ moved from SLURM to multi-cloud AI infra with @skypilot_org. ✅ 10x faster dev cycles ✅ SLURM-like convenience, K8s' reliability ✅ Scale on any infra blog.skypilot.co/abridge/

The American mind cannot comprehend driving for 23 hours and still be in Ontario

✨ @AltaMedHealthS, the nation’s largest FQHC, is partnering with Abridge to support more than 500,000 patients in Southern California with Abridge’s industry-leading multilingual capabilities and more. Read more about this groundbreaking partnership: abridge.com/press-release/…

Not now kitten daddy's experiments aren't going well

Removing query params before sharing a link is a basic sign of respect

(Pejoratively) Was this vibe coded?

Diagnosing the latest catastrophic mechanical failure of my 2009 Subaru Outback by sound alone

We’re so back

Nothing humbles the aspiring machine learning scientist faster

🎉 𝗔𝗻𝗻𝗼𝘂𝗻𝗰𝗲𝗺𝗲𝗻𝘁: Abridge welcomes Sagar Sanghvi, former Instacart CFO and partner at Accel, to fuel the next phase of growth as Chief Financial Officer. Watch the video to learn more about @spsanghvi4 and why he joined Abridge.