Joel Alexander (@joel_a_wilde) 's Twitter Profile
Joel Alexander

@joel_a_wilde

Building @PareaAI (YC S23) - LLM experimentation toolkit for teams collaborating with SME's

ID: 1219354093912842240

linkhttps://www.parea.ai calendar_today20-01-2020 20:20:53

376 Tweet

143 Followers

242 Following

Joel Alexander (@joel_a_wilde) 's Twitter Profile Photo

We've seen so many folks struggle with prompt optimization. What to change and why? Now, w/ OSS Zenbase all you need is example data and Zenbase will create an optimized prompt for you. 🔥

Joschka Braun (@joschkabraun) 's Twitter Profile Photo

Some tactics to build LLM apps consisting of multiple components: - test every sub-step to minimize cascading failures - reference-based evaluation of sub-components using synthetic data - cache LLM calls to quickly iterate on independent components More details incl. synthetic

Joschka Braun (@joschkabraun) 's Twitter Profile Photo

How do you detect unreliable behavior of your LLM app? Recently, we talked to the team at Sixfold and they shared with us a simple, yet powerful way to assess the reliability of their LLM app using PareaAI. More about how they test their risk assessment AI solution for

How do you detect unreliable behavior of your LLM app?

Recently, we talked to the team at <a href="/sixfoldai/">Sixfold</a>  and they shared with us a simple, yet powerful way to assess the reliability of their LLM app using <a href="/PareaAI/">PareaAI</a>. More about how they test their risk assessment AI solution for
Joel Alexander (@joel_a_wilde) 's Twitter Profile Photo

It’s crazy that a cars power is still defined by horse power. Like I have no mental model for the strength of a horse. In 18th century this probably made great sense,

Joel Alexander (@joel_a_wilde) 's Twitter Profile Photo

See so many open source AI repos written in typescript, for example all the CLIs are. Very interesting, is full stack typescript going to become the default GenAI stack ? Or all the good stuff still built with Python and just not open sourced?