Agentic coding assistants have transformed the way we build software – and they can be even more impactful for AI development itself. Instead of using LLM APIs, we can take back control and build our own AI systems, bootstrapped for us by agents. In this talk, I'll show you these new agentic development workflows, what they mean for developer tools and why code and the open-source ecosystem are more important than ever.
I'll also share some practical examples of how we built agents to help developers solve real-world NLP tasks and create practical and modular solutions that are small, fast and fully data-private. This requires human and agent collaboration: apps and developer APIs need to speak the same language, and humans need to be able to work alongside the LLM, not just with a model in the loop.
Read about and join the waitlist for Ellf, our new platform and virtual assistant for applied NLP and agentic NLP development.
https://explosion.ai/blog/against-llm-maximalism
A lot of people are building truly new things with LLMs, like wild interactive fiction experiences that weren’t possible before. But if you’re working on the same sort of NLP problems that businesses have been trying to solve for a long time, what’s the best way to use them?
https://speakerdeck.com/inesmontani/let-them-write-code-keynote-pycon-india-2019
Our philosophy for building developer tools and why good tooling needs good programmable interfaces.
Are we heading further into a black box era with larger and larger models, obscured behind APIs controlled by big tech monopolies? I don’t think so, and in this talk, I’ll show you why.
https://explosion.ai/blog/atomic-nlp
An applied NLP methodology inspired by Atomic Design: building reliable language understanding systems out of small, composable components instead of one big model and a prompt.
https://explosion.ai/blog/human-in-the-loop-distillation
This blog post presents practical solutions for using the latest state-of-the-art models in real-world applications and distilling their knowledge into smaller and faster components that you can run and maintain in-house.
https://speakerdeck.com/inesmontani/building-ai-with-ai
In this talk, I'll show you why we should use LLMs to build systems instead of as systems, and why code and the open-source ecosystem is more important than ever, not less.
inesmontani
hanhan1978
koukimiura
utilforever
16bitidol
starfish719
findyinc
mizchi
jsoizo
minodriven
.png){kind=avatar}
yuriko1211
gotok365
niftycorp
holman
aki_iinuma
vanstee
joshbly
takai
bkeepers
kneath
iamctodd
427marketing
mojombo
eileencodes
auna
