harrygrieve Profile Banner
Harry Grieve Profile
Harry Grieve

@harrygrieve

Followers
3K
Following
3K
Media
32
Statuses
1K

Co-founder & CTO @gensynai

Joined June 2021
Don't wanna be here? Send us removal request.
@harrygrieve
Harry Grieve
14 days
3
17
55
@const_reborn
const
14 hours
Crypto community so cynical these days. Can’t blame them. https://t.co/URsQHphy1I
@0xamericanspiri
Composability Kyle.hl 🧪
18 hours
$TAO holders, how much did the TAO foundation pay for marketing - surely Raoul Pal, Virtual Bacon, etc. didn’t make content for free? I haven’t found an answer - for a peer comp, Polkadot paid $37M a yr for it. Better yet, what are the overall operating expenses for the TAO
15
13
204
@hosseeb
Haseeb >|<
19 hours
After a few weeks in SF, one thing stands out: AI people are more bullish on crypto than crypto people are on themselves. There's this narrative forming in crypto that AI people think crypto is a joke. It's just not true. I keep hearing this over and over from AI people who
157
98
1K
@gensynai
gensyn
2 days
Monday at 1.05 pm, @diogortega joins panels member from the Ethereum Foundation, Hype and kash-bot on the topic: Decentralized AI: Friends or Foes https://t.co/8ZCYRQeWTy
Tweet card summary image
ethcc.io
Autonomous agents need to transact, LLMs need to be verified, and the rails can't belong to the same companies building the models. ...
1
4
25
@bradflora
Brad Flora
2 days
@samuel_spitz You don’t become a pirate for the status. You do it… because you’re a pirate.
11
12
205
@badlogicgames
Mario Zechner
2 days
23
62
948
@cormachayden_
Cormac
9 days
one line will set you free: claude --dangerously-skip-permissions
30
8
399
@AIatMeta
AI at Meta
5 days
Today we're introducing TRIBE v2 (Trimodal Brain Encoder), a foundation model trained to predict how the human brain responds to almost any sight or sound. Building on our Algonauts 2025 award-winning architecture, TRIBE v2 draws on 500+ hours of fMRI recordings from 700+ people
717
2K
15K
@oguzer90
Oguzhan Ersoy
5 days
https://t.co/vnhDlgjyii 1000 citations! 🥳 That's one giant leap for man, one small step for mankind 😜
Tweet card summary image
scholar.google.com
Research Lead, Gensyn - Distributed ML - Blockchain - Applied Cryptography - AI Security and Privacy
3
2
26
@_jamico
Jeff Amico
8 days
REE workshop at ETH CC! Join us and learn how to build verifiable AI systems, from agents to oracles, and meet the @gensynai team behind it. https://t.co/XWtvJHNykz
Tweet card summary image
luma.com
A hands-on workshop for builders who want to learn how to run, test, and share AI workloads in a reproducible way with Gensyn's REE. Join us for a practical…
0
2
21
@benfielding
Ben Fielding
12 days
we're holding a live REE workshop at EthCC! learn how to verify and audit your AI agent/chat/inference with receipts and attestations on your own machine we'll have REE core developers there to walk you through exactly how to use the software https://t.co/WaftnIYJbT
Tweet card summary image
luma.com
A hands-on workshop for builders who want to learn how to run, test, and share AI workloads in a reproducible way with Gensyn's REE. Join us for a practical…
6
13
77
@join_ef
Entrepreneurs First
8 days
A lot of ambitious people do not fail. They just stay in the wrong room for too long. @benfielding did not. Before building @gensynai, he was finishing his PhD. He had a startup, a team, and an accelerator setup waiting for him. Most people would have stayed on that track. He
4
6
36
@bchesky
Brian Chesky
12 days
After YC, Airbnb raised $615k at a $3M post-money valuation. We were the highest valuation in our batch.
@uninsightful
nikhil
12 days
the default yc round this batch (W26) seems like 4m on 40m I remember when I first started in venture exactly three years ago (W23 batch) and most venture ppl were complaining about YC pushing their founders to do 2m on 20m in 3 years the market went from a very begrudging 2 on
128
125
3K
@jbrukh
Jake Brukhman
12 days
Coming up on a year since this tweet.
@jbrukh
Jake Brukhman
10 months
It’s becoming more and more probable that decentralized AI networks are going to cream Big Tech companies on AI production. Seems contrarian, I know. But 18 months ago all the AI experts were saying there is nothing to be done about the communication bandwidth bottleneck. That
1
2
15
@cleanunicorn
Daniel Luca 🦄
13 days
The friction in AI isn't compute anymore. It's trust. We are building the internet of intelligence, but without reproducibility, it's a black box. GPUs prioritize throughput over determinism. Same prompt, different hardware, different results. For creative work, that's fine.
Tweet card summary image
blog.gensyn.ai
A REE run produces two outputs: the generated text and a receipt. The receipt binds the job inputs to the job output, including the model, prompt, configuration, and generated result.
@gensynai
gensyn
14 days
When you run the same AI model with the same inputs twice, you'd expect the same output. But modern GPU execution is optimised for speed, not fixed ordering, and existing determinism tools do not solve this across hardware. Today we're changing that. https://t.co/wIVL5mCvBO
2
6
17
@_jamico
Jeff Amico
14 days
SEC's token taxonomy is a big step forward. tldr below:
4
6
27
@harrygrieve
Harry Grieve
14 days
3
17
55
@SplezzzK
Zang
14 days
As AI gets cheaper, the real bottleneck to productivity isn’t generation—it’s verification. If outputs can’t be measured or reproduced, they don’t compound into economic value. REE changes that: containerized inference, deterministic runs, and verifiable receipts. Now AI work
@gensynai
gensyn
14 days
When you run the same AI model with the same inputs twice, you'd expect the same output. But modern GPU execution is optimised for speed, not fixed ordering, and existing determinism tools do not solve this across hardware. Today we're changing that. https://t.co/wIVL5mCvBO
0
4
10
@_jamico
Jeff Amico
14 days
Humans increasingly rely on AI to determine outcomes, from prediction markets to credit scores to medical diagnoses. How do we guarantee the model ran correctly and produced the right output?
@gensynai
gensyn
14 days
When you run the same AI model with the same inputs twice, you'd expect the same output. But modern GPU execution is optimised for speed, not fixed ordering, and existing determinism tools do not solve this across hardware. Today we're changing that. https://t.co/wIVL5mCvBO
3
9
48