DSPy
3.3K posts
@DSPyOSS
An open-source declarative framework for building modular AI software. Programming—not prompting—LLMs via higher-level abstractions & optimizers.
I will share two works at #ICML2026! Shared Lexical Task Representations Explain Behavioral Variability In LLMs > Thu, Jul 9, 5 – 6:45 PM (Hall A) Machine Studying: A System-Level Reframing of Continual Adaptation from Declarative Corpora > Fri, Jul 10, 3:30 – 5 PM (Hall A)
SITUATION EXPLAINED: What is token capital efficiency? We asked @kmad, AI Engineer and Developer "Instead of token maxing and using Opus and GPT 5.5 for everything, can you have a strategy as a business to measure the effectiveness of your spend on tokens? I call this token capital efficiency." "The way I would argue you do that is you inventory your tasks, you prove out what works, what is automatable, what you want to automate. You define your evals for measuring how successful that will be." "Once you have that kind of flywheel, then you can walk down that cost curve. 'Here's my acceptable level of performance. I want 97% accuracy or better on whatever this task is.' Once you can measure that, then you can try a bunch of different models under the hood." "What is that optimal point between retaining your level of performance and the lowest cost possible?"
My RLM agent can effortlessly process ~80k lines of service logs from CloudWatch in a single go. that's worth like 8 million tokens. The cool part is, after 53 steps, it had spent only 32k "active" tokens* (not through the full 8MM yet atp, more like half). That's nothing for Claude Fable 5 (rip), and weeell within effective context window, so its very "context-efficient". It can go VERY far and I dont even have to handhold it or anything, i'm not worrying about context running out or compactions either. I'm saying I kicked this thing off, almost without any context, and it was able to infer the service architecture based on logs alone, and spot issues my team didn't. In this particular case it was able to narrow down on a specific slice and find a couple issues that flew under the team's radar (AgentCore's throttles, Slack's user_not_found) Very handy. I'll release this as OSS soon (my first release on llm tooling!)
Continual learning is widely discussed right now, but mostly as improving on the job or avoiding catastrophic forgetting. But it has a different, difficult, and already urgent form: Given nothing but a corpus of documents, how should AI systems develop expertise in a new, unfamiliar domain? We call this problem Machine Studying.
Continual learning is widely discussed right now, but mostly as improving on the job or avoiding catastrophic forgetting. But it has a different, difficult, and already urgent form: Given nothing but a corpus of documents, how should AI systems develop expertise in a new, unfamiliar domain? We call this problem Machine Studying.
