"Generating thousands or millions of trajectories will help us find the right mechanisms for incorporating feedback and re-training those LLMs [Large Language Models]. There are numerous opportunities for developing new training engines, learning engines, and ultimately creating new benchmarks. We cannot improve what we cannot measure. The benchmarks are critical for assessing our current position."
Quote Details
Added by wikiquote-import-bot
Unverified quote
0 likes
Computer scientists from the United StatesStanford University facultyCalifornia Institute of Technology alumni
Original Language: English
Available Languages (1)
Sources
George Lawton, "Why better AI needs better simulators", diginomica, September 10, 2025
https://en.wikiquote.org/wiki/Silvio_Savarese
Revision History
No revisions have been submitted for this quote.
Categories
Silvio Savarese
7 quotes on TrueQuotesView all quotes by Silvio Savarese →
Related Quotes
"If you want to build a service agent that does returns, you cannot afford for the returns to be wrong one out of ten …"
"These are very similar to what DeepMind did when they were training AlphaGo. They have instructions on how to play Go…"
"This doesn't cover all of the objects, fields, and complexity of the orgs that use Salesforce. But even within this s…"
"Prompt engineering is a bit of an art. So, there's a lot of work that goes into that, but finding an automatic way to…"
"Our goal is to scale it up and make it as close as possible to real-world use cases. I think it's actually the next s…"
"This kind of scenario can be implemented by plugging it into the right structure, such as CRM-Arena, where we have st…"
"Americans have always been especially prone to regard all things as resulting from the free choice of a free will. Pr…"
"Democracy is clearly most appropriate for countries which enjoy an economic surplus and least appropriate for countri…"
"Here, for the last time together, appeared a triumvirate of old men, relics of a golden age, who still towered like g…"
"Rule 1 of cryptanalysis: check for plaintext."