theGupta @RealHGupta - Twitter Profile

28 days ago

Cool paper from PwC. "Earlier is always better" is the default intuition for agent clarification. New paper claims that's mostly wrong. Goal clarification loses nearly all of its value after just 10% of execution. The team built a forced-injection framework that drops ground-truth clarifications at controlled points along a long-horizon agent's trajectory, across 4 information dimensions (goal, input, constraint, context), 3 benchmarks, and 4 frontier models. 84 task variants, 6,000+ runs. Pass@3 falls from 0.78 back to baseline. Input clarification keeps value through roughly 50%. Past mid-trajectory, asking any clarification at all performs worse than never asking. A complementary study of 300 unscripted sessions shows no current frontier model asks within the empirically optimal window. 52% of sessions over-ask. Others never ask at all. Why it matters: clarification has been treated as a binary capability, does the agent ask or not. This is the first quantitative demand curve for *when* the question is worth asking. Paper: https://t.co/U4prpHjKgP Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

dair_ai's tweet photo. Cool paper from PwC.

"Earlier is always better" is the default intuition for agent clarification. New paper claims that's mostly wrong.

Goal clarification loses nearly all of its value after just 10% of execution.

The team built a forced-injection framework that drops ground-truth clarifications at controlled points along a long-horizon agent's trajectory, across 4 information dimensions (goal, input, constraint, context), 3 benchmarks, and 4 frontier models. 84 task variants, 6,000+ runs.

Pass@3 falls from 0.78 back to baseline. Input clarification keeps value through roughly 50%. Past mid-trajectory, asking any clarification at all performs worse than never asking.

A complementary study of 300 unscripted sessions shows no current frontier model asks within the empirically optimal window. 52% of sessions over-ask. Others never ask at all.

Why it matters: clarification has been treated as a binary capability, does the agent ask or not. This is the first quantitative demand curve for *when* the question is worth asking.

Paper: https://t.co/U4prpHjKgP

Learn to build effective AI agents in our academy: https://t.co/LRnpZN7L4c

9

127

23

101

12K

theGupta @RealHGupta

about 1 year ago

@grok tell me why there is a problem between North Indian and South Indian in terms of their local language?

1

0

12

theGupta @RealHGupta

over 1 year ago

@TheDeshBhakt @KIITUniversity @kpsharmaoli Who gives them the right to do this?

0

1

0

59

theGupta @RealHGupta

almost 2 years ago

@pbillore141 Bhaut sahi kam kiya tha aapne bhai..

0

4

Who to follow

G | 🇮🇳

@itsstaxy

Jai Jagannath🙏 | | Jai Hind🇮🇳 || RCB

Himanshu Mohanty🚀

@CodingHima

I make ideas into reality. Founder : https://t.co/jUOt2FzMzB

Simply Sujal

@362Sujal

Building https://t.co/0sjgKp2sN4 & https://t.co/LakR6qKhz7

theGupta @RealHGupta

almost 2 years ago

Full of excitement .. #INDvsSA

0

62

theGupta @RealHGupta

almost 2 years ago

Lots of hope was from @imVkohli

0

14

RealHGupta retweeted

Himanshu Kumar

@codewithimanshu

almost 2 years ago

500 TB Tutorials+ Books+ Courses+ Trainings+ Workshops+ Data Scientists Kit 🔥 Free Giveaway💖 ▪️ Data science ▪️ Web Dev ▪️ AI ▪️ Cloud ▪️ BIG DATA ▪️ Data Analytics & more. To get it : 1. Must follow @codewithimanshu to get the Link 2. Like and Repost 3. Comment "send"

195

154

90

19

8K

theGupta @RealHGupta

almost 2 years ago

@codewithimanshu Send

0

14

theGupta @RealHGupta

almost 2 years ago

@codewithimanshu send

0

5

RealHGupta retweeted

AISoC Society @AISoC_Society

over 3 years ago

Deep dive into the world of AI and play and test your skills with guessing prompts, creating stories, drawing objects, using hand gestures and many more to control and play games! Registration at: https://t.co/mG6pckn5bW For more information, visit: https://t.co/4DonJdp71H