@deepqlearning
 Zurich, Switzerland
Joined March 2025
  • Tweets27
  • Following423
  • Followers61
8Media
Robert Müller retweeted
Binfeng Xu
@billxbfSep 21
Today we give Superintelligence back to its owners. Introducing Skill2Env 🎨 , the most aligned and diverse dataset to fuel modern Agentic RL . 🧵
4
17
7
183
66,842
Robert Müller retweeted
Xiaomi MiMo
@XiaomiMiMoSep 21
We’re open-sourcing Pro and Flash, MiMo-V2.6-Distill-Qwen-9B, the technical report, 7K+ RL task environments, an end-to-end RL framework and composable mini-harnesses. Reproduce, verify and build on the work.
9
32
19
422
44,296
Robert Müller retweeted
Robert Jakob
@robertjakobJul 17
We are launching Aionic Labs (@aionic_labs). 🔥🚀 With support from SPRIN-D’s Next Frontier AI initiative, we are building a frontier AI company at the heart of Europe’s leading AI ecosystem. Clinicians decide when a patient needs intervention. Factory teams catch machine problems before production stops. Operators determine which alarms demand action. These are some of the most consequential decisions in the physical economy, and they all depend on time-series data. We are developing Time-Series Language Models (TSLMs) that can understand these signals, reason over them, predict what comes next, and suggest actions in natural language. TSLMs connect sensor data with language experts use to make critical decisions. Europe is the right place to build this technology. We have world-class research, top technical talent, and decades of industrial expertise and data. Our team brings together former founders, frontier tech experience, and world-class researchers from Stanford, Harvard, ETH Agentic Systems Lab (@ETH_agent_lab), ETH AI Center (@ETH_AI_Center) and CDTM (@cdtm_munich). We have built companies, developed technology at scale, and advanced research at the frontier of AI. Over the coming months, we will: → Launch selected TSLM pilots across manufacturing, healthcare, energy, logistics, telecom, insurance, and banking. → Build an open research community around OpenTSLM (@OpenTSLM) and TimeNet. → Enable trusted data partnerships across industries. → Accelerate the Time-Series AI community in Zurich and Europe. From Zurich. For Europe. 🇨🇭🇪🇺 Join our mission and help build a Frontier AI Lab from day one. Career and pilot program links are in the comments. Managing Directors: Robert Jakob (@robertjakob), Thomas Kaar (@thomaskaar) Supported by: SPRIN-D (@SPRIND) Photo: Felix Adler für SPRIND
25
26
7
110
10,636
All this AI and we're still finding sublets in group chats like it's 2009. So @_balsam0 and i built clawby. Text it what you want, it finds your match and intros you. Both sides say yes before any details are shared. live and free: clawby.chat/ 🦀
1
115
Excited to share that members of our lab co-authored 6 papers accepted at #ICML2026, including three Main Track and three Workshop papers 🔥🚀 📄 Accepted papers: ▪️ OpenTSLM: Time-Series Language Models for Reasoning over Multivariate Medical Text- and Time-Series Data [Main Track] ▪️ Auditing Emotion-Vector-Steered Political Bias in Open-Weight LLMs [AI4GOOD Workshop] ▪️ Reinforcement Learning of Karma Bidding Strategies [NExT-Game Workshop] ▪️ Cinematic Source Separation with Dialogue-Driven Sidechain Ducking [Workshop on Machine Learning for Audio] ▪️ Reinforcement Learning for Tool-Calling Agents in Fast Healthcare Interoperability Resources (FHIR) [Main Track] ▪️ Preserving Plasticity in Continual Learning via Dynamical Isometry [Main Track] @robertjakob @PatrickLanger20 @gaborhollbeck @f14wn @DerRiehl @atoof_sh @deepqlearning @kev_osull @cs06thegreat @nzuma0 @maxrosenblattl Huge congratulations to everyone involved. We are looking forward to presenting these works, reconnecting with colleagues, and meeting new friends in Seoul 🇰🇷 🔗 Full links in the comments.
5
6
1
20
2,255
The lesson learned is that agent evaluation needs environments where capabilities compose under pressure. Memory alone is not enough. Math alone is not enough. Tool use alone is not enough. The question is whether the policy remains coherent when the world pushes back.
1
6
195
Static benchmarks measure what a model knows. Interactive benchmarks measure what a model keeps doing when its own actions change the state of the world. Cattle Trade is intended to be a first step in that direction. Paper: arxiv.org/abs/2605.14537 @ETH_agent_lab @robertjakob
1
5
191
This is only the starting point. Obvious extensions:... -Frontier models at full reasoning budget -Memory and scratchpad ablations -Allow agents to chat -Comparison against humans Using the traces not just to rank agents, but to train them via self-play and RL
1
4
104
Why this matters: Agents for procurement, negotiation, sales, portfolio management, auctions, scheduling, resource allocation. None of them solve one-shot questions. They act in interactive systems where opponents react, incentives conflict, and state-tracking errors compound.
1
5
99
Different incentives for bidding spirals: NEED: price out the auctioneer to close your quartet EXTRACT: auctioneer pays you via buy-right BLOCK: win a card to stop a rival's quartet BID UP: make a rival pay the inflated price PUNT: speculatively win a card nobody collects yet
1
3
96
more replies
LLMs had many of the basic pieces. They could bid, trade, bluff, follow rules, complete sets. They didn't reliably compose these abilities into a stable policy when other agents reacted and the game state changed after every move.
1
3
85
EXTRACT is the most cognitively demanding spiral. You're not trying to win the card. You're modeling the auctioneer's willingness to keep it, pushing past their threshold so they pay you to walk away. Gemini 3 Flash and Gemini 3.1 Flash Lite EXTRACT most and perform best.
1
3
107
Gemini 2.5 Flash Lite challenged for a goose with zero money cards. Opponent countered with 10 coins and won for free. Gemini 3 Flash paid far above face value for a fourth quartet. Multiplicative scoring made it worth +1,800 points.
1
3
113
What separates winners from losers is strategic coherence: spending efficiency, resource discipline, phase-adapted bidding, composed together. Top agents end games broke with 4 quartets. Cash hoarders flatline at 2. Holding money isn't the goal. Converting it is.
2
4
78
Strong agents shift gears across the game. Gemini 3 Flash preserves cash early and ramps bid intensity nearly 10x by late game, timed to quartet-completion pressure. Gemini 2.5 Flash Lite bids aggressively from turn one and stays flat. It skips the accumulation phase entirely.
1
2
98
Most LLM benchmarks ask: can the model answer the question? We ask whether a model stays coherent across 60 turns of bidding, bluffing, bargaining, hidden offers, liquidity management, and adversarial play. Inspired by the German card game Kuhhandel. Strategy under imperfect info
1
3
129
more replies
The strongest result isn't "LLMs are bad at games." Broad reasoning scores don't measure strategic integration. Same 7 LLMs, three benchmarks, three orderings. Gemini 3.1 Flash Lite: 2nd on Cattle Trade, 3rd on Chatbot Arena, last on AAI Index. GPT-5.4 Nano runs opposite.
1
4
121
A few hundred lines of code. No pretraining. No language. No "reasoning." Just structured state tracking, cash estimates, simple bidding rules. Across 242 games, it outranked 6 of 7 cost-efficient LLMs on TrueSkill. It doesn't forget who has money or what it's bidding against.
1
3
77
Public cards. Hidden money offers. Zero-value bluff cards. No change for overpayment. Buy-rights in auctions. Multiplicative scoring. You need memory, arithmetic, opponent modeling, and timing. All at once.
1
5
94
Why do models that score well on standard benchmarks still make bad strategic decisions? In Cattle Trade, accepted at @MALGAI_ICLR2026 , LLMs play a 4-player 50+ turn game of auctions, hidden offers, bluffing, and limited cash. Some LLMs lose to simple Python agents. 🧵
6
3
18
537
Load more