up to something prev. yc, pioneer, uchicago
Joined January 2010
- Tweets2.1K
- Following299
- Followers928
- Likes8.1K
This was awesome to help work on. It’s obvious to us that making agentic financial operations more reliable is huge. Feel free to reach out, would love feedback on the benchmark!
We're launching Bank Arena: 52 long-running banking tasks across several banking apps.
We gave frontier computer-use agents a bank to run, from recovering transfers to settling merchant payouts. The median task takes experts ~75 minutes to complete, and the highest-scoring models completed <31% of tasks.
Here's what we found ↓
Tshepo Mohlala retweeted
We're launching Bank Arena: 52 long-running banking tasks across several banking apps.
We gave frontier computer-use agents a bank to run, from recovering transfers to settling merchant payouts. The median task takes experts ~75 minutes to complete, and the highest-scoring models completed <31% of tasks.
Here's what we found ↓
Tshepo Mohlala retweeted
extremely excited to announce our $30m raise for Catalyst.
we are looking for incredible engineers to explore the future of finance as we enter the dawn of superintelligence
please reach out if this is you!
Introducing Catalyst.
What do you believe? How would you trade on it?
You notice data centers are popping-up everywhere, gold will skyrocket, or just know that Labubus will go out of fashion.
Catalyst researches your ideas, builds the optimal strategy, monitors events, and executes trades. Autonomously.
Tshepo Mohlala retweeted
been saying this for a while.
every good company will look like the quant shops.
the quant shops look like that because they operate in markets dominated by machines.
soon, all of us will operate in markets dominated by machines.
The Navier Stokes proof drama is good foreshadowing for what happens / becomes important in the entire economy going forward: spotting opportunities, pointing models at them, and validating solutions faster than others. In other words, the economy / knowledge discovery increasingly looks like quant trading / pod shops / HFs, where having a differentiated view of why an opportunity is valuable and/or realizing that an opportunity is valuable (and that it can be plausibly captured/exploited) faster than others is all that matters.
This is from a piece I wrote in October 2025 and I think it will continue to hold a lot of explanatory power.
Tshepo Mohlala retweeted
All I'm asking is for tech people who say they believe crazy shit to actually act crazy
Like I'm excited about technology. I don't think the singularity has begun. But if i did, I sure as shit wouldn't spend my final days creating shareholder value. I sure as shit wouldn't expect anything in my stock portfolio to hold any value at all after the techno-rapture
Shut the fuck up about the singularity unless you are going to actually live like you mean it
Tshepo Mohlala retweeted
Announcing Discovery Loop!
I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor.
♾
Learn more at: discoveryloop.com
You have to ask yourself very seriously if what you're working on is ambitious enough/worthy of your time during the singularity.
An internal version of Astra, @OpenAI’s next major model family, solved 10 major open problems in mathematics, quantum complexity, and theoretical computer science.
We believe it will be a major step for scientific reasoning. openai.com/index/ten-advance…
Tshepo Mohlala retweeted
We’re giving scientists, mathematicians, and engineers free access to our frontier models—starting with 10,000 researchers and expanding to 100,000 through 2027.
ChatGPT for Academic Researchers is built to accelerate discovery across disciplines.
Tshepo Mohlala retweeted
We've just made a major finding for SovereignAI 🚨:
Take any open-weight model (here: Qwen3.5-397B), apply our π-shaped Continual Learning, and own a genuine frontier model competitive with Opus 4.8 for ~$450k in compute.
Full tech report + open-source models coming soon!
I think if AI impact feels low in daily life, it’s because somehow it’s not superhuman at holistic strategic thinking yet. Many long horizon multivariate decisions it fails noticeably at. Maybe this is just a missing context problem. Maybe it’s an RL problem lol
Tshepo Mohlala retweeted
Building software by prompting agents without ever reading or editing the code directly isn't actually a new thing. It's just called being an engineering *manager*. We have decades of experience about how to be effective as engineering managers. It is a different skill, though.
.@Cloudflare mcp is great. Very happy I can manage dns, email rules etc. with agents now.
Tshepo Mohlala retweeted
Real brains follow Dale's principle: a neuron can either excite its neighbors or suppress them, but never both. Standard deep learning ignores this and uses backpropagation.
In our new paper, Diffusing Blame, we fix this disconnect. By introducing a routing method that broadcasts error signals directly to the hidden layers, we can train networks made of dedicated positive and negative neurons to strictly obey Dale's principle, all without relying on backprop!
This method works surprisingly well on image recognition tasks despite the strict biological constraints. We also achieved competitive, backprop-free reinforcement learning on complex locomotion tasks and the open-ended Craftax environment.
It is neat to see that representation learning remains possible even when we force deep learning to play by the rules of real neurons.
Introducing "Diffusing Blame": can a neural network learn competitively while strictly obeying Dale's principle, the rule that real neurons follow? We show it can, across both image classification and reinforcement learning. 🧠
Accepted at #ALIFE2026
arxiv.org/abs/2606.31700
Real neurons generally follow Dale’s principle: each neuron is predominantly excitatory or inhibitory. Standard artificial networks usually ignore this constraint, allowing every unit to mix positive and negative outgoing weights.
Backprop makes the gap even wider. Its backward pass needs exact transposed copies of the forward weights, the so-called "weight transport problem,” which biology doesn’t seem to have a mechanism for.
So we asked: can a network that strictly enforces Dale's principle still learn well, without weight transport?
Our approach builds on Error Diffusion (ED), a local rule that routes a single global error signal directly to every hidden unit, where each layer is split into separate excitatory and inhibitory streams with four non-negative weight matrices, so a synapse's sign comes from fixed population identity rather than a learnable weight. Our main contribution is to extend ED from binary to multi-class problems via modulo error routing.
We then asked whether this routing mechanism could provide useful credit signals in the noisy setting of reinforcement learning. During PPO training on Ant, Humanoid, and HalfCheetah, we compared each local ED update with the corresponding true backpropagation gradient. Among the routing schemes we tested, modulo routing consistently produced the strongest alignment.
Taken together, these results show that Dale-constrained networks can still learn without transporting weights backward, suggesting a potential path toward learning rules that are both effective and more biologically plausible.