Jamieson Warner (@Jobamey) profile picture

Trying to reproduce the developmental program behind brain connectivity. Website: jamiesonwarner.con

San Francisco, CA
Joined May 2015
Photo by Jamieson Warner (@Jobamey)Photo by Jamieson Warner (@Jobamey)Photo by Jamieson Warner (@Jobamey)Photo by Jamieson Warner (@Jobamey)Photo by Jamieson Warner (@Jobamey)
Shimmers of consciousness… They have brains too. It’s beautiful, isn’t it?
I love watching animals dream. When bad people say that animals don’t have souls, I wonder what they could possibly mean when it’s so obvious that they have their own inner experiences of the world, their own inner worlds and imaginations.
1
21
Has anybody noticed that like half the replies on this website are very obviously written by Grok? As a kid I used to make anonymous accounts and reply to people online and here it's the same but all the commenters are Grok behind an anonymous account.
1
24
It'll be cool! It'll be the future! They say. > Causes nasuea
1
12
Looking at *you*, electric cars and VR.
10
Cats are natural yogis.
1
22
Jamieson Warner retweeted
Interestingly, the gradient properties arising from shift invariance of softmax(teapawt.org/miyashita_03/status/21…) and scale invariance of a network(teapawt.org/miyashita_03/status/21…) can be derived using exactly the same steps. Both start from a symmetry: transforming the input or weights by a constant "c" leaves the function's output unchanged. Then differentiating with respect to "c" reveals the constraints on the gradient. - Shift invariance: Gradient components sum to zero. - Scale invariance: The gradient is orthogonal to the weights.
2
26
209
9,400
These next ~12 months sit at a golden opportunity before AI tools have replaced us as researchers entirely, but while our access to next-generation automated coding can vastly accelerate any idea you have and bring it to life.
1
35
Jamieson Warner retweeted
post AGI it’s just about finding peace and being hot
21
23
1
277
10,090
ES seems to work pretty well in high dimensional search environments. You can just solve for good operating reasoning behaviors in the region around the pre-trained model, and this lets you locate solutions to long-horizon task environments typical in RL.
Ok guys so all the recent weight-space programming buzz inspired me to write a blog post on zeroth-order, weight-space methods (back to the basics), from the theory to a bit of update on our latest result. Turns out plain weight perturbations don't just work for fine-tuning on math, they work for post-training agents too: vanilla ES takes Qwen3.5-4B from 12.1% → 21.1% on Terminal-Bench 2.1, past the released Tmax RL checkpoint (18.2%). Better yet, no credit assignment, no critic, no masking, no messy hyperparameters, just good ol' perturb-and-update. Oh did I forget to mention that all with population size of ~30? Here's why I think this could be huge for long-horizon agentic post-training 🧵+blog post👇
3
95
Humans will continue to commune with the super intelligence, who will understand our mental limitations and deliver a firehose of relatable concepts. The research work will be done by AI, but a select few “ascended” humans may choose to lose themselves in connection to it.
If we want to continue progress in science, we must be prepared to acknowledge that the correct description of nature will eventually simply be too complicated for humans to understand. I do not think it is wise to make human understanding a requirement. AI is the way forward.
49
Biology has been doing computer science for four billion years. Before the discovery of DNA, Alan Turing and John von Neumann worked out the logic of self-replicating machines, showing how matter folds into a system that reads and executes instructions to build itself.
13
46
8
390
19,516
Technical debt is now when you don’t stop to talk to your coding agent.
26
Jamieson Warner retweeted
reflecting on 1920s images of people working on massive generators for the electrification of the modern world makes me realize that the past was wildly futuristic too
126
505
58
6,772
159,276
I tend to agents as a day job
23
Science is fundamentally about tool creation and use. Design an experiment (a measuring tool) to measure a variable.
28
Jamieson Warner retweeted
83
4,778
241
57,005
946,632
Wandering through the dunes, call that Chaparral Roam
1
97
Less investment, more action
32
Better Data efficiency = Better Generalization = Better Parsimony (Occam’s razor)
1
38
Jamieson Warner retweeted
How do you build a brain from a genome? We know a lot about the mechanisms and molecules Here we revisit the algorithmic problem: What kinds of programs can specify a brain within the genome’s information budget and the finite time available for development?
11
71
6
350
25,928