gensyn (@gensynai) / X

gensyn

1,561 posts

gensyn

@gensynai

Networking the core resources for machine intelligence to flourish alongside human intelligence. Discord discord.gg/gensyn | Foundation @GensynFND

Joined April 2020

Pinned
gensyn
@gensynai
Jun 12
The Role of Feedback Alignment in Self-Distillation New research published by @SemihSmile and @oguzer90 "It concentrates the learning signal exactly on the tokens where reasoning breaks, and leaves correct steps intact. All from natural language" Semih explains below.
Semih Kara
@SemihSmile
Jun 12
Multi-agent LLM systems are everywhere, and agents talk to each other: critiquing, correcting, giving feedback. A design lever: what style of agent-to-agent communication works best? At @gensynai, we designed one. New style, same agents, better results. blog.gensyn.ai/the-role-of-fe…
8.1K
gensyn
@gensynai
14h
Gensyn is partnering with @tokensandai for the Multiagents Hackathon, London June 26 Join to build with Gensyn's REE and be in the running for prizes on the day More info and registration in the thread below👇
gensyn
@gensynai
18h
The @tokensandai Multiagents Hackathon is this Friday June 26 in London UK Join to build with Gensyn's Reproducible Execution Environment and be in the running for multiple prizes, with Gensyn team members on site throughout the day More details, links and registration👇
5.1K
gensyn
@gensynai
18h
The @tokensandai Multiagents Hackathon is this Friday June 26 in London UK Join to build with Gensyn's Reproducible Execution Environment and be in the running for multiple prizes, with Gensyn team members on site throughout the day More details, links and registration👇
tokens&
@tokensandai
Jun 23
We’re heading to london! 🇬🇧 🏆 Winners get up to 6 months of Cursor Ultra. Cash prizes. 10,000+ in credits. Come build with us this Friday alongside teams and prizes from @cursor_ai @GoogleDeepMind @ElevenLabs @tavilyai @gensynai Prometheux @twilio @tessl_io @GoSenso and
8.5K
gensyn
@gensynai
18h
2/ Read more on REE and watch the video tutorial: • docs.gensyn.ai/tech • github.com/gensyn-ai/ree • youtu.be/fkXGAWxNSSk
00:00
648
gensyn
@gensynai
18h
3/ Register for the Multiagents Hackathon on luma:
Multiagents Hackathon · Luma
From luma.com
530
gensyn
@gensynai
Jun 23
"If you do self-distillation with step-aligned feedback,..., that amounts to process supervision, but you don't need to train a reward model" - @SemihSmile on newly published research: The Role of Feedback Alignment in Self-Distillation Read the article based on his talk here
gensyn
@gensynai
Jun 22
Article
The Art of Telling a Model It’s Wrong
Setting up the problem A multi-agent workflow looks straightforward until it has to be wired up. In multi-agent workflows there are at least two models in conversation: a smaller one doing the work,...
7.1K
gensyn
@gensynai
Jun 23
The Role of Feedback Alignment in Self-Distillation has also been accepted for presentation at the @icmlconf 2026 RLxF Workshop on July 10 Find the publication on @arxiv here
arxiv.org
The Role of Feedback Alignment in Self-Distillation
Conditioning a language model on additional context, such as feedback on a previous attempt, typically improves its response. Self-distillation trains the model to retain this improvement when the...
682
gensyn
@gensynai
Jun 23
Read more in @SemihSmile's thread and the Gensyn blog article x.com/SemihSmile/sta…
Semih Kara
@SemihSmile
Jun 12
Multi-agent LLM systems are everywhere, and agents talk to each other: critiquing, correcting, giving feedback. A design lever: what style of agent-to-agent communication works best? At @gensynai, we designed one. New style, same agents, better results. blog.gensyn.ai/the-role-of-fe…
The Role of Feedback Alignment in Self-Distillation
From blog.gensyn.ai
758
gensyn reposted
Oguzhan Ersoy
@oguzer90
Jun 23
Replying to @benfielding @lior_eth and 4 others
TLDR: - Sakana Fugu: learned multi-round orchestration hidden behind one model API. - OpenRouter Fusion: configurable parallel panel, judge, and final synthesis pipeline. More detailed: Sakana Fugu is more like a learned, autonomous, adaptive orchestration model. You call one
742
gensyn reposted
Hervé Cuviliez - Id4.vc
@hervecuviliez
Jun 21
Decentralized compute isn’t a sci-fi concept anymore—it’s actively becoming the infrastructure layout for the next generation of AI. @gensynai was right all along. With @Tesla latest release, plus @AMD and @nvidia pushing hard into "compute at home" setups, the tech giants are
Polymarket
@Polymarket
Jun 20
JUST IN: Tesla has filed a trademark application for “MEGAPOD,” described as modular data center hardware for AI computing.
Gensyn | the network for machine intelligence
From gensyn.ai
1.6K
gensyn
@gensynai
Jun 22
Article
The Art of Telling a Model It’s Wrong
Setting up the problem A multi-agent workflow looks straightforward until it has to be wired up. In multi-agent workflows there are at least two models in conversation: a smaller one doing the work,...
8.4K
gensyn reposted
Jeff Amico
@_jamico
Jun 19
The way to keep open source legal is to make it even more open. Weights, data, recipe should all be open source and provable. Show you didn't train on harmful data or steer the model. Doesn't eliminate risks but strikes the right balance.
Nathan Lambert
@natolambert
Jun 19
Banning open-source AI in any form would be a mistake. A general audience PSA with @kevinsxu on why open source upholds American values. Managing frontier risks is hard, but reducing transparency, innovation, and education from kneecapping the open frontier would be worse.
4.4K
gensyn
@gensynai
Jun 18
AI Market Validation is live on @Delphi_fyi
Delphi
@Delphi_fyi
Jun 18
NEW UPDATES! AI Market Validation is live! When a market is created, AI jumps in and helps the market creator improve it. A little helping hand from a friend 🤝 @jud_bonneville explains below 👇
4.1K
gensyn reposted
Judson Bonneville
@jud_bonneville
Jun 18
1/ AI market validation is live on @Delphi_fyi this process sanity-checks every stage of the market creation process: the question, outcomes, sources, settlement prompt, and market timing, then flags issues for you to fix before you publish now available for whitelisted market
Delphi Marketplace Waitlist
From gensyn.typeform.com
2.6K
gensyn
@gensynai
Jun 18
Article
The Missing Dispatcher for Expert Models
Most people reach for a large generalist model as if it were a Swiss army knife. One tool answers a casual question, works through a derivative, and rewrites a function, and it does each of them well...
2.9K
gensyn
@gensynai
Jun 18
Read more on the Gensyn blog DEI: Diversity in Evolutionary Inference for Quality-Diversity Search
Christian Catalini
@ccatalini
Jun 17
Replying to @gensynai
It intuitively makes sense and reminds me of one of the most cited papers in organizational science (March 1991). I wouldn’t be surprised if the approach proves particularly important for business/strategy/market applications. jstor.org/stable/2634940
DEI: Diversity in Evolutionary Inference for Quality-Diversity Search
From blog.gensyn.ai
2.8K
gensyn
@gensynai
Jun 17
gensyn
Host
The Role of Feedback Alignment in Self-Distillation
33 tuned in
Jun 17
31:29
1.5K