“We're actually able to bias the model based off of the direct user feedback using some of these like online training methodologies. We're able to see like real upticks in user satisfaction, completion rates, decreased tool call failures”

Unsupervised Learning with Jacob Effron · Ep 94: Applied Compute CEO on the Limits of RL, the New AI Hyperscaler & Why Post-Training Wins Inference

Play from 16:09
16:09

Audio streams from the publisher’s own feed. Verbatim does not host or copy it.

What was described

bias a model using direct user feedback from an assistant product, applying online training methods to increase user satisfaction, completion rates and reduce tool-call failures

Said nearby on the same subject

“A collective SaaS product … community … you and your other SMEs in the background … the user feedback loop that you can create in that scenario.”
NC
Nick Coleman
Lawyerist Podcast ·
Why it worked Share this moment
“We’re not training our models … we’re using customer feedback to make our systems better.”
NC
Nick Coleman
Lawyerist Podcast ·
Data play Share this moment
“A separate copy of the model that’s running on the tool calls to judge whether it’s a safe thing to do… or if it is safe, is it what the user asked for”
AA
Andrew Ambrosino
The Stack Overflow Podcast ·
Automation Share this moment
“We provide an end user interface so that users can select from multiple model providers… administrators have complete governance and control over how data is treated before it ever leaves their network boundary”
♫
Said on Using AI at Work: AI in the Workplace & Generative AI for Business Leaders
Why it worked Share this moment

Moments like this, every Monday

The strongest new moments on your topics, in the speakers' own words. Free for three topics.

No card. No password. One-click unsubscribe.

Track a competitor?

Pro watches up to 25 names or phrases and emails you the morning after they are said.

See Pro, $29 a month

Verbatim quotes a short extract and links to the episode; the episode belongs to its publisher. Remove a show.