Udayan Umapathi (@udayan2025) header
Udayan Umapathi (@udayan2025) profile picture

Founder and CEO @voltalabsinc to build bio-automation from the ground up. Before: experimental physics, design, art, electronics @medialab @CypressSemi.

Cambridge, MA
Joined October 2009
Photo by Udayan Umapathi (@udayan2025)Photo by Udayan Umapathi (@udayan2025)Photo by Udayan Umapathi (@udayan2025)Photo by Udayan Umapathi (@udayan2025)Photo by Udayan Umapathi (@udayan2025)Photo by Udayan Umapathi (@udayan2025)Photo by Udayan Umapathi (@udayan2025)
Udayan Umapathi retweeted
Thank you to everyone who packed the room for our #AMP2026 workshop on automating long-read nanopore workflows in AML 🙏 Keep the conversation going at booth #721. Expo hall open today until 7 p.m. PT. voltalabs.com/contact
3
3
516
Udayan Umapathi retweeted
Great to run into our AMP speaker Dr. Parth Shah (Dartmouth Health) at the gate ✈️ Volta Labs and Parth, flying to Seattle together for #AMP2026! Workshop tomorrow, Wed 11/11, 4 PM PT. Booth #721. voltalabs.com/amp-2026
1
2
95
I've been an early user of speech interfaces for years, and for the last two I've been building speech tech on the side and surrounded by people working on silent speech and BCI. Everyone is racing to build better sensors. I don't think sensors are the bottleneck. Data is. Speech is one spectrum: out loud, whispered, mouthed, subvocal, BCI. Every step right is more private, but the signal is weaker and there's less data. Silent speech will work for strangers when we train it on speech said out loud, not by collecting hours of silent data from e
1
1
2
185
Training data is where it gets hard. Speech out loud has a huge corpus. Move right and you have a fraction of that, and you need hundreds of hours. Ground truth is hard too. If someone mouths a word, how do you know what they meant? Out loud, you just listen.
1
27
Generalization is the hardest part. Different accents and languages, different faces and throats, different ways of mouthing the same word. Out loud, clean data in the major languages already exists. To the right, you collect it yourself, one person at a time.
26
This is more or less how I work at Volta. Although I haven't gotten to build my own harness.
I've changed how I work over the last few weeks. I now have one lead agent for each area of focus, running all the time, and a chief of staff agent that oversees them. Each lead manages its own sub-agents that write code, test it, chase reviews, and coordinate with team members. Everything I need to decide or review comes into one prioritized queue, a simple local web app. Each product review includes screenshots and a video demo of the feature. I clear the queue in batches, mostly by voice and keyboard shortcuts. The leads also bring back ideas I hadn't considered. Working from one queue, instead of switching between dozens of chats, has cut a lot of mental overhead. The system keeps learning my preferences and distilling my judgment into a "Brian brain," so its recommendations and first drafts get better before they reach me. The goal is fewer corrections per change. Every change still goes through the same human review and security checks. It is eliminating tasks, not human judgment. What an exciting time to push the frontier of productivity, especially as we build the economy for agents (agentic finance). We should all keep pushing in this direction.
104
We’ve reached a stage where, with the help of @bot, we can generate almost bug-free code for hardware development. However, issues arise when deploying on actual hardware, and these problems intensify when fluids are manipulated by hardware, and even more so when biology is involved. Validation guidelines are crucial for building a fully automated development pipeline that follows the sequence: 1) Software, 2) Hardware, 3) Fluids (Physics), and 4) Biology.
Replying to @poteto
Without a strict set of validation guidelines agreed upon inside the org, it’s quite clear how that goes (poorly). Critical infra, F500, etc. cant “just automerge thousands of PRs” yet. I don’t doubt you are doing that on some kinds of software. But I bet that software doesn’t have human lives or livelihoods on the line. If it does, I’d say that w/ current tech that’s irresponsible.
1
134
hey @elonmusk @bot and @poteto is there a tier above Ultra for heavy users? I've maxed out on the Grok Bot usage, and cost jumped a lot past the Ultra limit. Would love a higher usage plan or the option to pick a cheaper model for routine tasks, so my usage lasts longer. This could also help grow revenue without blocking heavy users. For context, my work spans all the way from sales / marketing to building financial models to developing algorithms for computer vision.
Replying to @DogecoinNorway
The cost of serving Grok Bot was unsustainably high relative to revenue
123
He was ahead of his time
One reason I am sanguine about the technological singularity is that I have seen it coming for over 50 years. In 1975, when I was 17, I wrote: "Before man can most effectively aid his problem solving machinery, he must first understand it. At present, his understanding is very poor. However, when the breakthrough in understanding the brain occurs, improvement will come very fast and in an exponential increase. Once man has created an intelligence greater than his own, that intelligence should be able to create an intelligence greater than itself, as man whose intelligence is less than its was able to create its intelligence. Thus, the level of intelligence will shoot up with incredible rapidity, probably within the space of years. Within a few years of a breakthrough, which may occur any time now, man will either have been left behind by a superior intelligence of his own making, which he can no longer understand, or he will have become that intelligence and will no longer be like anything imaginable today."
109
Labs should drive themselves, like autonomous cars. @kirk3gaard got Callisto from the box to sequencing in about four hours. No calibration, no automation engineer, just load and walk away. That's what the autonomous lab looks like, and it starts with sample prep.
Four hours from unboxing to sequencing. Rasmus Kirkegaard at Aalborg University took Callisto from install to a nanopore sequencing run the same afternoon. No calibration, no automation engineer. Load and walk away. Labs should drive themselves, like cars.
1
2
892
Hey @bot and @poteto, some feature requests for @bot's interface from a heavy user. I've got bots managing other bots: a chief of staff and a program manager handing work out to specialist bots for finance, sales, BD, and more. It's genuinely changed how I run the company. But working with every bot as its own chat doesn't scale. I'd love to pop bots out into their own windows so I can watch and steer several at once, like a control room instead of an inbox. Same goes for docs and spreadsheets. The speed is incredible. A bot will rebuild a cash model or revise a term sheet in minutes. But then I download the file, open it, review it, and type feedback back into chat. That loop is the bottleneck now. What I want is to look at the sheet or doc right next to the bot, point at a cell or paragraph, and say "fix this." The bots are fast enough. The interface just needs to keep up.
2
1
145
Happy to be your design partner on this. Give me early access and I'll build the interface I wish existed.
32
Udayan Umapathi retweeted
Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel. - Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all. - Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract. - Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code. - And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while. The funnel is driven by a combination of factors: 1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel. 2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel. 3. Access. Free-tier, paid-tier, internal. So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.
Judging by my tl there is a growing gap in understanding of AI capability. The first issue I think is around recency and tier of use. I think a lot of people tried the free tier of ChatGPT somewhere last year and allowed it to inform their views on AI a little too much. This is a group of reactions laughing at various quirks of the models, hallucinations, etc. Yes I also saw the viral videos of OpenAI's Advanced Voice mode fumbling simple queries like "should I drive or walk to the carwash". The thing is that these free and old/deprecated models don't reflect the capability in the latest round of state of the art agentic models of this year, especially OpenAI Codex and Claude Code. But that brings me to the second issue. Even if people paid $200/month to use the state of the art models, a lot of the capabilities are relatively "peaky" in highly technical areas. Typical queries around search, writing, advice, etc. are *not* the domain that has made the most noticeable and dramatic strides in capability. Partly, this is due to the technical details of reinforcement learning and its use of verifiable rewards. But partly, it's also because these use cases are not sufficiently prioritized by the companies in their hillclimbing because they don't lead to as much $$$ value. The goldmines are elsewhere, and the focus comes along. So that brings me to the second group of people, who *both* 1) pay for and use the state of the art frontier agentic models (OpenAI Codex / Claude Code) and 2) do so professionally in technical domains like programming, math and research. This group of people is subject to the highest amount of "AI Psychosis" because the recent improvements in these domains as of this year have been nothing short of staggering. When you hand a computer terminal to one of these models, you can now watch them melt programming problems that you'd normally expect to take days/weeks of work. It's this second group of people that assigns a much greater gravity to the capabilities, their slope, and various cyber-related repercussions. TLDR the people in these two groups are speaking past each other. It really is simultaneously the case that OpenAI's free and I think slightly orphaned (?) "Advanced Voice Mode" will fumble the dumbest questions in your Instagram's reels and *at the same time*, OpenAI's highest-tier and paid Codex model will go off for 1 hour to coherently restructure an entire code base, or find and exploit vulnerabilities in computer systems. This part really works and has made dramatic strides because 2 properties: 1) these domains offer explicit reward functions that are verifiable meaning they are easily amenable to reinforcement learning training (e.g. unit tests passed yes or no, in contrast to writing, which is much harder to explicitly judge), but also 2) they are a lot more valuable in b2b settings, meaning that the biggest fraction of the team is focused on improving them. So here we are.
2,266
5,941
2,296
44,362
7,013,389
Udayan Umapathi retweeted
That’s it!!! 🙌 Mega job from the team, it’s been amazing all weekend. And now we keep going @redbullracing!
615
7,705
235
57,671
691,758
Rasmus and the Aalborg team went from setup to a fully loaded P24 of peatland soil samples. Looking forward to digging into how they did it, live on Oct 29. Bring your questions 👇 @voltalabsinc @nanopore @kirk3gaard
Thanks @kirk3gaard and the Aalborg team! 🌱 Rasmus joins our CEO live Thu Oct 29 (10:00 ET / 15:00 CET) to talk Callisto setup, ONT ligation prep and soil metagenomics. Free, with live Q&A 👇 airmeet.com/e/aea93610-c260-…
1
117