Joined August 2013
- Tweets 3,793
- Following 1,059
- Followers 4,026
- Likes 2,008
299 Photos and videos
Pinned Tweet
Feb 13
New followers.
No researcher is to be found here. I am a novel capabilities engineer doing random walk.
(read: unhinged)
1
10
1,657
Not a lawyer-nerd.
If a frontier model company, materially changes a prompt a user provided are they accepting exposure proportionate to the causal contribution and any duty assumed? Especially if you've voluntarily undertaken to do that, and the voluntary-undertaking doctrine, especially if its marketed for safety, since intervening to reduce harm is exactly what creates the assumed duty?
Seems like, intuitively at least, modify users input is a liability landmine.
5
339
Jun 13
Jun 13
The US government, citing national security authorities, has issued an export control directive to suspend all access to Fable 5 and Mythos 5 by any foreign national, whether inside or outside the United States, including foreign national Anthropic employees.
The net effect of this order is that we must abruptly disable Fable 5 and Mythos 5 for all our customers to ensure compliance.
Access to all other Claude models is not affected.
We apologize for this disruption to our customers. We believe this is a misunderstanding and are working to restore access as soon as possible.
Read our full statement: anthropic.com/news/fable-myt…
4
169
Adam Hibble retweeted
Jun 11
I still feel this way, 8 years later.
28 Mar 2018
People sometimes ask if I think it's risky for everyone to have access to AI. I think it's MORE risky for an exclusive & homogeneous group alone to develop tech that impacts us all. x.com/techreview/status/9787…
2
16
113
12,706
Jun 10
I really don't like the feeling of anger man. It sucks.
Im really disappointed by anthropic. I've been writing model code since 2010, everything I do looks like 'frontier research'.
Like genuinely. The whole day has felt so shitty.
8
8
212
24,986
May 26
I have been thinking about the measure of how far away a transform moves an input from its 'automorphism group'.
Intuitively, this would seem to specialise (in expectation) to negentropy / expected log-det Jacobian.
vaguly, it'd result in flows that act more as a gaussanity tests/regulariser, where the 'goal' is to starve the flow, ie isometry log-det -> 0, singular values -> 1
1
5
613
May 26
This is cool because you could use flows with expensive inverses like MAFs since you’re throwing them away after training
284
May 15
gonna revisit this, i really liked the split head formulation for compute reuse.
18 Dec 2025
A Ladder of Heads Varient of Free Transformer
Each transformer layer contains a binary VAE: where the "Encoder" and "Decoder" live inside the same Transformer layers, separated only by attention masks and head splits. Decently efficient thanks to flex attention.
First N heads: bidirectional (encoder) — sees full context for posterior
Remaining heads: causal (decoder) — computes prior from z_{<i}
Gives you an single forward pass encoder / decoder with shared weights and a bunch of computational reuse. The prior is conditioned both on the causal state, but also implicitly on the previous posterior/prior samples.
Still tweaking beta/free bits and gotta do a non toy train etc. But I think its a pretty neat formulation.
github.com/Algomancer/Ladder…
9
1,151
May 15
artifact from a discord chat
wake-sleep Bayesian program learning in one PyTorch file. Integer arithmetic domain. Joint inference over a typed contextual PCFG and a neural recognition model that amortizes the posterior over programs given I/O examples.
basically a nano dream coder
github.com/Algomancer/bayesi…
5
323
May 10
quietly good paper for anyone training discrete latent variable models. lower variance gradients for discrete latents, with a nice reframing of the whole estimator family. this one rao-blackwellizes the noisy bit and patches the leftover bias with a control variate.
Did a quick implimentation here and a mapper layer for vaes like free transformer.
github.com/Algomancer/Better…
2
3
38
3,662
May 10
me: pull out this into a single .py for github
claude: got it im changing all your numerics for no reason
4
286
May 1
sparsemax core. tropical-pilled. λ-maxxing from hard max to mean. real zeros fr. fused kernel no thoughts head empty. vaguely trains. idk had needs
github.com/Algomancer/moreau…
1
9
480
Mar 12
Okay this in unhinged.
But you can add a hook, that fakes being a agent team mate and make your application interface be on an agent team.
gist.github.com/Algomancer/0…
2
7
637
Mar 12
Combine with this, x.com/Algomancer/status/2031…
and have fun.
Mar 10
there is something really magical about giving an agent access to the dom directly
setup a js term, embed it, open claude, give it an api it can curl to to execute js - pretty fun.
minify an app to a single file, insert it on the fly via curl.
I think if this becomes like, less of a security nightmare its totally a thing.
3
254
Mar 10
there is something really magical about giving an agent access to the dom directly
setup a js term, embed it, open claude, give it an api it can curl to to execute js - pretty fun.
minify an app to a single file, insert it on the fly via curl.
I think if this becomes like, less of a security nightmare its totally a thing.
10
899
Mar 4
This is my new favourite way to start a project.
Self-bootstrapping Claude app. Three files, one loop. Tell the agent to modify itself. It will.
github.com/Algomancer/claude…
1
6
377
Mar 2
I said this a while ago, but dude im so keen for code conditioned realtime video generation. We are close.
react.js? Nah, just bash pipe to pixel space.
30 Sep 2023
I think not enough people are thinking about what high quality, long context video generative models mean in the longer term. It's not just generative Netflix.
You're going to have generative video applications on your generative operating system, everything you see on your screen can be generated. Including chat interfaces.
1
9
524