INNOVATION - AI Dynamics

Ideas for Improving Language Models Without New Developments

By

–

28 March 2023 19h27

reply if you have other ideas about how language models could currently be improved without any new developments and I'll add them to the thread

→ View original post on X — @alexalbert__

28 March 2023

Language Models Gaining Access to Tools and APIs

By

@alexalbert__

–

28 March 2023 19h27

giving language models access to tools and api’s chatgpt has done this with plug-ins, no longer can you say chatgpt is bad at math or doesn't have access to current info soon every api on the internet will be forced to have language model interaction documentation

→ View original post on X — @alexalbert__

28 March 2023

Chained prompt sequences as the future of AI APIs

By

@alexalbert__

–

28 March 2023 19h27

using pre-created chained-prompt sequences instead of just one single prompt as api costs continue to decrease this will become more popular, the intermediate prompts will be hidden and only the output will be shown could imagine chatgpt integrating a "prompt marketplace"

→ View original post on X — @alexalbert__

28 March 2023

Unlocking Language Models’ Potential Without New Development

By

@alexalbert__

–

28 March 2023 19h27

gpt-5 is not needed to 100x the potential these models have we could stop all language model development today and we still haven’t scratched the surface of their capabilities here are a few non-obvious ways language models can be improved without creating any new models:

→ View original post on X — @alexalbert__

28 March 2023

Parallel Language Models Orchestrated by Conductor LLM

By

@alexalbert__

–

28 March 2023 19h27

running language models in parallel with each one focused on a sub-task, all orchestrated by a conductor language model picture something like a massive tree of GPT models working on answering a single complex prompt

→ View original post on X — @alexalbert__

28 March 2023

Local Language Models Using Cloud Models as Plugins

By

@alexalbert__

–

28 March 2023 19h27

language models running locally on your machine that treat more complex language models in the cloud like plug-ins

→ View original post on X — @alexalbert__

28 March 2023

Talent matters more than formal education in tech

By

@jessyseonoob

–

28 March 2023 19h12

Pour moi c'est : t'as des gens doué, et des gens pas doués. Qu'ils aient fait des études ou pas, et y'en a pleins qui apprennent sur le tas qui sont encore plus doués. Y'en a qui ont fait des etudes et qui sont pas doués. c'est histoire de personne et de competence pour le bon

→ Voir le post original sur X — @jessyseonoob

28 March 2023

Hectare: Type-Directed Program Synthesizer 100x Faster

By

@mit_csail

–

28 March 2023 18h15

Want to get a program from just its type? Meet new type-directed synthesizer Hectare. Over 100x faster than state of the art on some problems — all with 10x less code. All thanks to a new kind of solver, ECTAs, solving a decades-old problem in synthesis: https://
bit.ly/3z2D5Ly

→ View original post on X — @mit_csail

28 March 2023