
cs @uwaterloo, prev @datacurve @menloresearch | 🏃♂️🫀 ✝️
🇸🇬/🇨🇦/🇺🇸
Joined June 2021
- Tweets1.7K
- Following1.1K
- Followers979
- Likes13.6K
hosting an event with @floguo @charlesxjyang @anjazehfuss and alan gou on the meaning of work and ambition in the age of AI on 1 Dec!!
in person for folks in Waterloo (location TBD), speakers will be calling in :D
This was inspired by events organized by @ACTS17org and the Vocare Collective over the summer. As LLMs become increasingly great at what they do, how should one approach thinking about their career? What are we ultimately working towards?
1
82
All university rankings are very flawed.
I wondered how we could rank unis as purely as possible? So I looked at the career path of ~100,000 engineers over 10yrs. Universities rank high when its graduates work at companies, ranked PageRank style, that engineers compete harder to join.
Stanford, MIT, Harvard were top 3, but the extent they were above the rest surprised me. Brown at #6, Waterloo at #7!
Daniel Ching retweeted
Asimov 1 is an open-source humanoid robot you can build, train, and customize.
- 1.2 m tall, 35 kg
- 25 powered joints
- Open hardware with off-the-shelf parts
- Open-source SDK and simulation models
Available as a DIY kit.
Order yours at menlo.ai/order
Daniel Ching retweeted
one of my favorite projects from this summer :)
Everyone thinks computer use agents are the future in healthcare, but you can't just let agents run amok in EHRs without any guardrails.
Our summer intern, @shayaan_azeem, took a better approach: he created a system for agents to build EHR browser automations without ever letting one act unsupervised in prod.
Shayaan was an awesome intern this summer, and it was a blast watching him build this.
Here's the full write-up: build.forus.com/ehr-browser-…
Daniel Ching retweeted
in April, we published a launch essay for @tacitlabsco
that outlined a thesis for frontier AI research in biology -
1. if we want models to be useful for curing disease, we must evaluate them on tasks with roughly that shape and complexity. optimize directly on the thing you care about - i.e. discovering & developing a drug
2. we should aim to model the biotech (process, reasoning, decision-making), rather than the underlying biology. biology is an empirical science - develop better empirical methods rather than treating it as an analytical domain
3. the most difficult part of all of this will be verification of those empirical methods with signal that actually matters (does the proposed drug actually work in a human?). biological time is a tough wall to run up against (most expensive forward-pass of all time). figuring out how to turn scientific problems into learning problems is a first-class research problem
4. if we believe (1) and (2), and we do a bunch of good work and figure out some clever solutions to (3) then we unlock increasingly agentic discovery/development, with the logical extreme being this concept of the "zero-person biotech", which is now (somehow) percolating in mainstream discourse
5. (4) has extremely exciting implications for the overall economics and practicalities of how we develop novel therapeutics (broadening viable aperture to rare disease, n-of-1 trials; new mechanisms of distribution) all of which are net positive for the world. much to be optimistic on here
discovering and developing a drug is one of the most complex and long-horizon tasks in the broad spectrum of human endeavors. inarguably one of the most important and high-impact research directions in AI today
Happening now!! Outside DP until 230
inspired by @askcliffe, we're hosting waterloo's very own give me an answer!!
pull up outside DP on Wed Oct 28th :)) first of a bunch of events 👀👀
I trained a 750K-parameter world-model agent on my M5 Pro to play video games!
Presenting NanoDreamer: a tiny Dreamer-inspired agent that learns game dynamics and trains an actor–critic in latent imagination.
Write-up: benjamin-bai.com/projects/na…
so encouraging to see tech twitter go off with truths that make us sing :D
This Friday, I’ll be giving a talk on the historical case for the resurrection of Christ.
Several years ago, I began looking into the historical evidence as an atheist to prove that Christianity was nothing but a therapeutic superstition. What I found completely changed my life.
Over the course of a year, I spent nearly every waking moment reading whatever I could find on the subject. My entire life was suddenly reorganized around answering this one question: did Jesus Christ actually rise from the dead?
There was a time when I thought this question was too ridiculous to even consider. But looking back, it was by far the most important question I've ever asked.
If you're interested in learning more, check out the link to the event page below. All are welcome to join regardless of religious background or lack thereof.
1
13
624
nice
I've been looking into public benchmarks and am astounded by the poor quality vs. how much we rely on them as an industry.
@xeophon pointed me to this excellent site epoch.ai/benchmarks/ which is an independent third party which scrutinizes benchmarks and "looks at the data" 💪
For example, here are some popular benchmarks we all rely on. Before trusting a benchmark its worth running them and inspecting examples the grader failed carefully, to get a sense of the quality.
I'm continuing to look more into this, and will write something up eventually once I've collected more data, but I think this is an important subject.
3
304
inspired by @askcliffe, we're hosting waterloo's very own give me an answer!!
pull up outside DP on Wed Oct 28th :)) first of a bunch of events 👀👀
Daniel Ching retweeted
What still limits a sub-2 marathon? And could a second fuel move it?
I wrote a technical analysis on running economy, durability and exogenous lactate. It has more questions than answers, on purpose. 🧵
Daniel Ching retweeted
i get asked VERY frequently to chat about data / RL envs / new companies coming into the space. all of the questions tend to be exactly the same so i want to summarize my thoughts:
Q: "Should I start a data company?"
A: probably not, unless you have some sort of alpha or have seen what frontier RL envs are like. basically, if you're not deep within an industry, seen data via a lab or working at another data company, or very well connected with lab researchers, you probably should not. though, if its your passion, now is the best time and it'll get increasingly worse. data is also much harder now than it was 6 months ago, and quality standards are ever-increasing.
Q: "Is there a product in the space?"
A: i think there are repeatable systems that can be created. i.e. QA systems, faster task generation, infrastructure around rollouts/VMs/GPU utilization/prod. system emulation etc. i think data is inherently bespoke, luckily LLMs allow you to harness bespoke intellect infinitely. i don't think it's a SaaS style product where it serves everyone immediately, but more of a quant-desk type system with alpha on how to do specific tasks embedded within an engineer-system like manner. each eng will have a desk that serves them specifically that unlocks 10x, 100x, 1000x capabilities.
Q: "Is {insert set} data valuable"
A: depends on what you do with it. raw data dumps are cool but it's a lot more about what you can turn it into.
Q: "How does {insert data company} differ from {insert competitor}?"
OR
"Should I join {insert data company}?"
A: if you're in the "big, legacy" data companies with contractors, it doesn't really differ that much. if i had to pick, i'd probably join mercor or surge from the big contractor companies, by virtue of early employees/founders being extremely smart. however, i'd much rather join a small, research-heavy data company. they are by far the companies with the highest slope.
Q: "Won't labs just in-house this anyway?"
A: (1) data is an additive game that isn't zero-sum. just because labs are increasing synthetic data creation internally does not mean they won't buy more data. especially given it is an arms race to the top of intellect, if you can just make valuable data that moves the needle, it doesn't really matter if they can replicate it, it's not worth their researcher time if you solve a problem. the caveat being you have to actually add something yourself, and not just be a parrot who takes requests. saving researcher time is the name of the game.
(2) i think this logic applies to everything, if you adopt this mentality, you probably should not work at any applied AI company, since they'll come after or absorb most products in the space.
*taps sign*
Heard today from an AI data entrepreneur:
"Contractors are gaming the system that data companies have developed. They're just spinning up synthetic pipelines themselves.
When I was a student, I did contracting for ML tasks with a data company. I set up a system where I was just using Gemini 2.5 pro to complete all the tasks, and they never noticed."
The provenance and quality of data has never been more important.
7
348
Daniel Ching retweeted
Biology lags code by about a year. But there are strong + unintuitive technical parallels.







































