Weekly Dose of Optimism #188

Curation edition covering AI-driven Alzheimer’s research, AlphaEvolve chip lithography breakthroughs, laundry-folding robots, Artemis II milestone, and education.

Curated Topics

1. OpenAI Foundation commits $100M to Alzheimer’s — Six research grants including Arc Institute’s “AI lab-in-the-loop” approach: perturb brain organoids, measure results, feed into AI models, iterate. Also funds David Baker’s Institute for Protein Design and EvE Bio.
2. Substrate x Google DeepMind AlphaEvolve — GDM’s evolutionary coding agent rewrote Substrate’s lithography simulation stack: 6.8x faster, 74% less memory, 97% lower cloud cost. Enables single-exposure patterning at 24nm pitch (2nm-node territory). American competitor to ASML’s EUV.
3. Arena Physica releases Heaviside — First EM foundation model from the company featured in the Electromagnetism essay. Playable in RF Studio.
4. Syncere Lume robot lamp — $2,499 floor lamps with six-axis arms that fold laundry autonomously. Uses VLMs for perception. Constrained-problem robotics approach.
5. Artemis II — Crew rounds Moon, sets human distance record at 252,756 miles from Earth.
6. Primer education — nbc portfolio company launches new brand campaign emphasizing resilience and mastery.

RDCO-Relevant Flags

• AI agents in science: Arc Institute’s lab-in-the-loop is the clearest example yet of autonomous AI agents driving physical-world research loops — directly relevant to our agent architecture thinking
• AI for chip design: AlphaEvolve’s 97% compute cost reduction on lithography simulation is a concrete case study for “AI accelerating its own hardware” feedback loop
• Data engineering: Substrate’s computational lithography stack is essentially a data pipeline optimization problem — the kind of work RDCO could consult on