Al Merose

Yo, space is old hat. We’re naming AI models after elements now blog.google/innovation-a...

Gemini 4 Argon: our next era of frontier intelligence blog.google

Sneak peek: I implemented a 3 layer MLP that predicts the fashion MNIST dataset using SQL (with my own autograd impl inside the database). Apparently einsum notation is better expressed as a JOIN. github.com/xqlsystems/x...

xarray-sql/benchmarks/nn.py at claude/xarray-sql-mnist-demo · xqlsystems/xarray-sql github.com

Maybe a troll, but also maybe a union one day: concernedaistaff.org

Coalition of Concerned AI Staff concernedaistaff.org

New result: Automatic derivation of FlashAttention from vanilla attention in ~25 lines of SQL via three portable extensions. github.com/xqlsystems/x...

xarray-sql/benchmarks/ml/01_attention.py at 6debe2495dcb4e8fac20ab97df5b89baae2a31a6 · xqlsystems/xarray-sql github.com

Not yet merged, but I added autograd to my array database experiment. I’m building a new NN framework primarily in SQL. github.com/xqlsystems/x...

Differentiable SQL: grad()/jvp()/vjp() over xarray and any DataFusion table by alxmrs · Pull Request #192 · xqlsystems/xarray-sql github.com

With my friend @realcromwellian.bsky.social’s help (and his Mac Studio), we got a 126M param NanoGPT model training in ~165s — primarily written in SQL. github.com/xqlsystems/x...

xarray-sql/benchmarks/ml/02_nanogpt.py at main · xqlsystems/xarray-sql github.com

In other words, you’re free to write the mathematical definition of attention in SQL, and you get the high performance just by adding on my smart query planner: github.com/xqlsystems/e...

GitHub - xqlsystems/einfold: Fast tensor contractions in SQL github.com

I would use this tree map from USGS as a ground truth to train a small regression model on top of Alpha Earth Embeddings to get a close bound. data.fs.usda.gov/geodata/rast... deepmind.google/blog/alphaea... Then, I would add data to see if I could make the bound tighter.

AlphaEarth Foundations helps map our planet in unprecedented detail — Google DeepMind deepmind.google

Oy vey. Better to avoid that direction. I’m constraining myself to a $20/mo plan for my side projects. Instead, I’m trying this with my sub agents: github.com/xqlsystems/e...

Sequence agent work around the review frontier by alxmrs · Pull Request #26 · xqlsystems/einfold github.com

All of our code is open source, and contributions are welcome: github.com/m2lines/Samu... We're working on fitting our model training into a (free-tier) colab notebook, so anyone can help advance climate science: github.com/m2lines/Samu...

GitHub - m2lines/Samudra github.com

All of our code is open source, and contributions are welcome: github.com/m2lines/Samu... We're working on fitting our model training into a (free-tier) colab notebook, so anyone can help advance climate science: github.com/m2lines/Samu...

Colab quickstart by YuanYuan98 · Pull Request #722 · m2lines/Samudra github.com

Banger paper alert: This uses the P=NP complexity theory to attempt to answer a question about gravity, using the difficulty of the NP-hard problem to argue that gravity must be quantum. What a cool proof by contradiction! arxiv.org/html/2606.14...

Semiclassical Gravity Efficiently Solves 𝖭𝖯-Complete Problems arxiv.org

For file type, there is at least one type of social media built around it: images > instagram, video > youtube, vertical video > tiktok, audio > clubhouse, safetensors > huggingface, executable files > github

I’m like 6 months to a year behind using all sorts of agentic coding patterns (I just asked Claude to delegate impl to subagents, etc.) and honestly (surprisingly?), I think I’m better off for it.

Agentic engineering feels like “problem engineering” to me. Coding becomes defining goals and constraints for agents to optimize towards. Like all coding, this requires taste and discernment.

Rust is ideal for agentic programming because large compile times give you an opportunity to think and reflect

Open source project idea: give that old phone in your drawer new life by installing Prism ML’s LLM on it to make it a stateful agent.

2004: real graphics programming is hard, it can’t just be simple C programs. 2011: real machine learning is done in C or C++, Python is just a toy or pedagogical language. Now: SQL for machine learning is cursed and will never work.

I don’t understand why there is so much agentic tooling. I feel very productive with just Claude equipped with the gh CLI, using standard GH features like issues and comments. While I use prompt patterns, I feel most cases require sufficient nuance that I don’t want to capture something in a skill.

I’m happy to be working on open source science and language models because the biggest safety risks of AI, imo, stem from concentrations of power — in corporations, governments, and their intersection.

If there is one contribution that I hope to make to computer science, it’s that relational algebra is as important as linear algebra or calculus is to machine learning.

The norm for ML in Python is to script at the intersection of at least three layers of programming language (Python, C++, then the CUDA stack). So, what if we could replace this with just Python and SQL + the freedom to choose your own physical layer?