Harshit Joshi

@harshitj__

CS phd @StanfordNLP, @StanfordOVAL | prev: @MSFTResearch | LLM systems for knowledge access, discovery and curation

Stanford, CA

harshitjoshi.com

Joined May 2012

426 Following

2,457 Followers

Replies Reposts Media

Pinned
Harshit Joshi
@harshitj__
Oct 15
Scientific 🔬 progress often starts with an unexpected observation that sparks a new hypothesis. But AI agents often start with a hypothesis, then look for observations to support it. We believe agents should actively explore both observations and hypotheses. When observations
00:00
4
Harshit Joshi
@harshitj__
Oct 9
although suggesting the next message could be useful, it might also limit users thinking and lead everyone to converge on the same analyses, experiments, and results. I believe research thrives on diversity, with everyone trying different things -- really hope we don't lose that
OpenAI Developers
@OpenAIDevs
Oct 9
Now in beta: composer predictions in Codex for Pro users. Codex can now suggest your next message based on your conversation and how you talk to it. One of the most loved new features we've ever tested internally.
00:00
Harshit Joshi
@harshitj__
Jul 10
llms are literally horrible for generating football worldcup trivia questions and the options w/o search then I asked it to be funny and wth is this (funny for sure):
1
Harshit Joshi
@harshitj__
Jul 1
would be nice if I didn't have to monitor my claude code sessions and had a shepherd!
Weiyan Shi
@shi_weiyan
Jul 1
Ever watch an agent solve a problem with 100% confidence and get it completely wrong? You can fix it… if you catch it. But then you need to watch every… single… run. New paper: We show that a meta-agent can do this for you, so your agents get ~2x more done on their own. 🧵
00:00
2
Harshit Joshi
@harshitj__
Jun 18
Had so much fun competing and WINNING with @ShichengGLiu @PriyankShethia and @JiudingSun. Special thanks to @AnthropicAI — our sponsor and mentor for the event! Also Databricks team is very goated for hosting such an amazing competition and coming up with this benchmark!
Databricks
@databricks
Jun 18
Stanford University is the 2026 Databricks Grounded Reasoning Cup champion! 🏆 The inaugural Grounded Reasoning Cup at #DataAISummit 2026 is in the books, with Stanford taking 1st, UMass Amherst earning 2nd, and Yale University securing 3rd after six high-intensity rounds of
2