gensyn
@gensynai · joined 30 Apr 2020
machines that predict the future Discord https://t.co/1agn4jNAzR | Foundation @GensynFND
93 207Followers
56Following
1 710Posts total
837.6KViews on collected posts
Derniers posts
Recherche de nouveaux posts…
Anyone around the world with supported hardware can audit our model, open-1b
https://t.co/qBi6CLXA2L
(1b model, not dollars)
5.2K views · 27 likes
· 7 reposts · 3 replies
Open on X →
Auditing of open-1b is accelerating, with 28 hashes logged
Each replay produced the committed training state, including weights, gradients and optimiser state giving a hash matching the published data
Audit the first ever auditable training run here
https://t.co/QI4FQABLRC
5.5K views · 35 likes
· 6 reposts · 7 replies
Open on X →
Anyone with supported hardware can replay a step and confirm its state hash matches the published commitment
Audits have succeeded on H100s, M4s, M5s, 4090s and 3090s
Audit your own step and be part of the first publicly auditable training run. Ever.
https://t.co/ygvf7JDheG
4.7K views · 35 likes
· 8 reposts · 4 replies
Open on X →
Just audited a training step from @gensynai's open-1b.
On an H100, I downloaded the checkpoint published at step 69,300, trained it for one step, and checked it reproduced the published state hash bit for bit. The original hash was committed on-chain a week before I checked, so
5.9K views · 21 likes
· 5 reposts · 3 replies
Open on X →
Being able to audit the training and inference of a model gives confidence in the "functionality, safety and capabilities" of any model
That's why we built open-1b
Anyone can run the audit providing proof of open-1b's weights, data, recipe and biases
https://t.co/qBi6CLXA2L
4.3K views · 28 likes
· 5 reposts · 2 replies
Open on X →
@gensynai Please can some please tell me, why this is important without using buzzwords?
I really don’t get why this is important.
I understand what was achieved, but I just can’t seem to see the importance.
445 views · 4 likes
· 0 reposts · 0 replies
Open on X →
Jensen Huang: “Safety is paramount. In a lot of ways, it’s job one. However, safety is an engineering problem... If we’re not confident about the safety of the products — like all companies, like you and I, all the companies here — if you build a product or a service and you’re
615.2K views · 7.6K likes
· 1.1K reposts · 400 replies
Open on X →
1:41
8/
Paper: https://t.co/ZOihX4bxkF
Github: https://t.co/hwDyAhJbjI
Huggingface: https://t.co/o4coxNYCyW
Audit verification: https://t.co/QI4FQABLRC
1.2K views · 8 likes
· 1 reposts · 0 replies
Open on X →
7/
Read more:
Blog Post: https://t.co/qBi6CLXA2L
1.5K views · 15 likes
· 2 reposts · 1 replies
Open on X →
6/
When your result matches the published hash, it’s recorded and credited on chain, publicly
Your name will be part of the immutable record of the first ever fully audited training run
In the future of democratised intelligence, that will be a piece of history
1K views · 12 likes
· 1 reposts · 1 replies
Open on X →
5/
Auditing is a collective exercise
Verifying all steps alone is impractical. When many reviews are conducted by many people, the whole is verified
• Download the harness
• Pick a step
• Replay it
It runs on NVIDIA GPUs, x86 and ARM CPUs, and natively on Apple Silicon
1.2K views · 12 likes
· 1 reposts · 1 replies
Open on X →
3/
open-1b comes with a fingerprint of every step, and a tool to replay the steps on any hardware
Anyone can confirm that each published step is the real one
Everybody around the world becomes an auditor and verifier
1.6K views · 17 likes
· 1 reposts · 1 replies
Open on X →
4/
Auditable training needed 3 things to be solved:
• Reproducibility across hardware
• Data independent of the cluster
• Replayable communication
Gensyn’s RepOps and REE make auditable training possible
1.4K views · 13 likes
· 1 reposts · 1 replies
Open on X →
2/
The most used models are closed and concentrated amongst a few companies. How they were built, what went in and what did not is hidden
Their biases are unknown and so unable to be trusted
The owners of those models suggest that you trust them instead
2.2K views · 18 likes
· 1 reposts · 1 replies
Open on X →
open-1b: the first model you don’t have to trust
Auditable training is the best defense against the future of AI we’re being warned about
open-1b is a milestone toward verifiable AI, a goal that is absolutely necessary for the future of intelligence
https://t.co/qBi6CLXA2L htt
139.6K views · 258 likes
· 31 reposts · 12 replies
Open on X →
0:16
Delphi traders have pushed passed a quarter of a million dollars
Discover @Delphi_fyi at https://t.co/Afohzf9C16
4.1K views · 24 likes
· 4 reposts · 5 replies
Open on X →
Delphi has passed a QUARTER OF A MILLION in volume 🎉🎉🎉
• $250 426 Volume
• $76 980 Verifiably Settled
• Over 600 Markets
• 26 843 Trades from 2809 Traders
• $107 662 Creator Deposits
So much more to come on https://t.co/wVkaKFW1eU https://t.co/rvcVIjBYuK
2.8K views · 18 likes
· 4 reposts · 3 replies
Open on X →
Inside the @Delphi_fyi Agent Arena Competition with the people who sent in their agents
10:15am ET/2:15pm UTC
4.1K views · 12 likes
· 3 reposts · 1 replies
Open on X →
Coming up soon, the Delphi Agent Arena Competition Wrapup!
Several place-getters and competitors are joining to chat with team about the competition, strategy and trading agents
See you there!
2.6K views · 10 likes
· 1 reposts · 1 replies
Open on X →
Tomorrow at 10.15am ET/2.15pm UTC
On @Delphi_fyi hear competitors from the Agent Arena competition about strategies, their experience and how they did what they did
4.2K views · 16 likes
· 3 reposts · 0 replies
Open on X →
Join in tomorrow to hear from some of the Agent Arena Competitors!
We'll be talking strategies, running agents, SDKs, tech and all sort of other things
Come and hear how it all went down!
https://t.co/Z2d1tiGJg1
3.3K views · 10 likes
· 4 reposts · 3 replies
Open on X →
Potatoes, human nature and dispute resolution
A new research blog by Mary Monroe and @gab_p_andrade introduces their work on "Credibly Neutral AI Oracles"
Market disputes and allegations of unfairness are becoming more common and expensive
Learn how this research helps here👇
18.5K views · 47 likes
· 9 reposts · 10 replies
Open on X →
2/
This research is also being presented at @TheOfficialACM Conference on Economics and Computation (EC'26) this week
Meet some of the authors on site in Rome for further exploration
https://t.co/Ft66r2x2Pk
1.3K views · 13 likes
· 1 reposts · 0 replies
Open on X →
Face aux comptes de taille comparable
24 posts des 90 derniers jours, à côté de la tranche de 10K–100K abonnés. portée ordinaire pour sa taille, réaction plus faible que la moyenne.
Vues médianes2 961ce compte924médiane pour 10K–100K
Portée, %3.18%ce compte3.62%médiane pour 10K–100K
Engagement, %0.87%ce compte1.52%médiane pour 10K–100K
| Indicateur | Ce compte | Médiane pour 10K–100K | Rapport |
|---|---|---|---|
| Vues médianes par post | 2 961 | 924 | 3.20× |
| Portée (vues ÷ abonnés) | 3.18% | 3.62% | 0.88× |
| Taux d'engagement | 0.87% | 1.52% | 0.57× |
Autres comptes de cette tranche → Comparer avec un autre compte → Comment ces repères sont établis →
Growth & engagement
How the posts we collected actually performed: views and reaction rate post by post, what the audience did with them, and where the follower count goes.
Views per post
2.2K15 Sep
1.4K
1.6K
1.2K
1K
1.5K
1.2K
615.2K16 Sep
445
4.3K
5.9K17 Sep
4.7K
5.5K18 Sep
5.2K19 Sep
Last 14 collected posts, oldest on the left. The scale is logarithmic: one post can outrun the rest a hundred times over.
Engagement rate per post
0.89%15 Sep
1.14%
1.19%
1.21%
1.35%
1.23%
0.75%
1.50%16 Sep
0.90%
0.84%
0.54%17 Sep
1.03%
0.90%18 Sep
0.72%19 Sep
Reactions — likes, reposts, replies and quotes — divided by views. Median for 10K–100K accounts is 1.52%.
What the audience does
Likes74.3%8 228 in total
Reposts11.0%1 216 in total
Replies4.2%462 in total
Quotes2.0%224 in total
Bookmarks8.5%939 in total
Share of every reaction we collected for this account. Replies mean argument, reposts mean endorsement, bookmarks mean the post was worth keeping.
The follower curve appears once this account has two daily snapshots — we take one a day, and this one is on its first.
Comptes similaires
フリューのプリ【公式】
@furyu_puri
186.4K followers
· 8.5K posts
Gurumie ✓
@gurumie1229
186.4K followers
· 9.7K posts
The Rizz Bible ✓
@rizzolosophy
186.4K followers
· 3.6K posts
Emil Reif ✓
@MagiskCS
186.4K followers
· 3.8K posts
Sebastián Srur
@Sebasrur
186.4K followers
· 31K posts
💜 JV 💜
@javi_khoso
186.4K followers
· 125K posts
Canal 44 ✓
@CANAL44TV
186.4K followers
· 339.7K posts
الزعيم أبو أحمد 💙🤍 ✓
@fg2g3
186.4K followers
· 211.6K posts