head of growth agnost ai (yc s26)
San Francisco, CA
Joined May 2025
- Tweets3.9K
- Following630
- Followers5.9K
- Likes3.4K
dhruvieiei retweeted
your AI assistant could be working perfectly on paper while frustrating real users.
that's exactly why we built Agnost.
find out where your AI is failing, why users are getting stuck, and what needs fixing.
try it at agnost.ai
quick walkthrough below!
why are companies like Google, Exa, and Corgi paying attention to what their AI assistants get wrong?
because just having an AI assistant that works isn't enough.
are users getting helpful answers? is the AI forgetting things? are people leaving frustrated?
made a quick video explaining how to find these problems before they hurt your product.
dhruvieiei retweeted
we've been everywhere on your timeline. now let us show you what we've actually been building haha.
meet Agnost AI.
we help companies figure out where their AI agents are messing up, why users are frustrated, and what needs to be fixed.
no more manually going through thousands of conversations.
here's a simple walkthrough of Agnost AI and how it works.
dhruvieiei retweeted
Looking for GPUs? Let us worry about it
We find you hidden GPUs in the market:
— Least lead time
— Best possible pricing
— Requirements of your choice
We have supplies of almost all kinds of GPUs, including H100, B300, A100, and L40S, for variable contracts. We have DC providers across the US, EU, India, and Southeast Asia.
DM me with details of your GPU specs, and we'll try to fit you in. Spammers and people testing the waters, please take a bad seat. We'll ask you to make a $9 transaction to start the discussion.
We have supplies ready to deploy
gpumafia.com
dhruvieiei retweeted
Introducing OpenAI’s Dots x Orthogonal
Try it out: orthogonal.com
Ask for a pipeline before bed. Wake up to verified leads, cited sources and the first replies.
Your always-on dots now have every tool they need, from company and people data to email.
dhruvieiei retweeted
🚨 Introducing official Agnost AI SDK for Mastra agents.
Just one line of change, and you get to see all your agent insights right in Agnost.
Docs link is in the comments.
dhruvieiei retweeted
IF YOU’RE STILL COPY PASTING BETWEEN CLAUDE, CODEX AND GROK, YOU’RE NGMI
today we are launching October to fix this
give October the goal, and it checks:
> which harnesses you have
> remaining usage
> what worked well for similar tasks before
then it picks the team, connects them, splits the work and lets them coordinate directly. e.g. codex with 20l 5.6 can lead, fable can review and grok can execute
and the whole thing is multiplayer 🎃
dhruvieiei retweeted
THE INBOUND ON THE PRODUCT SIGNUPS IS INSANEEEE!!!
WE'VE HIT THE HIGHEST PER DAY EVENT COUNT ALREADY! 3 MORE HOURS, SUPPORT US ON PH.
AGNOST AI (YC S26) JUST LAUNCHED ON PRODUCT HUNT!
AI agents don’t fail like traditional software.
They return 200. They pass evals. And they still quietly frustrate & churn users.
Agnost AI finds those silent failures that evals miss & only show up in production conversations.
Then we turn those failures into harness improvements & better models.
dhruvieiei retweeted
Today we’re launching @decawork backed by @ycombinator : one place for IT to deploy, govern, and maintain the internal AI agents running across a company.
Every team is building AI tools/agents now. They go live with real access, no standard way to ship, nobody watching.
With Decawork, every agent - built in Claude Code, Codex, or anywhere else - ships with its own identity and scoped credentials. IT signs off before it's live, sensitive actions wait for approval, and every run is watched from first to last.
I built AI at NVIDIA, deployed at OpenAI and Meta. Aman (@arxneo14) built Barclays' AI compliance platform used by internal teams, then shipped agents to 100K+ monthly users. We lived this and quit to solve it.
Internal AI agents/tools piling up at your company? Chat with us founders[at]decawork[dot]ai or decawork.ai
dhruvieiei retweeted
Tried @poppysimplified today.
It invented things about me I never said, then sent a reminder late.
Small misses, but for a personal assistant, memory + timing are basically the product.
Promising idea. Trust just needs to be tighter.
