17 papers
Never Stop Thinking: Continuous-Time Language Agents
Bojie Li, Noah Shi
Voice agents built on LLMs follow a rigid listen-think-speak loop that inserts seconds of dead air before every reply. We show that continuous-time cognition (thinking while listen…
The Trade-off Was in the Labels: Causal Supervision for Turn-Aware Streaming ASR
Bojie Li, Noah Shi
A voice agent must decide, moment to moment, whether the user has finished; silence rarely settles it: a caller reading a phone number pauses mid-digits, a one-word "Stop!" ends a…
The Model Knows Your Project, Not You: Measuring Recognition in LLMs with NameRank
Bojie Li, Noah Shi
What a frontier model recalls about a person or tool from its own weights -- before any retrieval step -- often shapes the first description a human sees, making that parametric co…
Interaction Scaling: Grounding the Third Axis of Test-Time Compute
Bojie Li, Noah Shi
There are two standard ways to spend more compute at test time: let a model reason longer, or sample more attempts and keep one. Both share a hidden limit: they are internal. Every…
Parametric Multimodal User Memory: Storing What Captions Cannot Carry
Bojie Li, Noah Shi
A personalized agent needs a user memory: a persistent model of who its user is. Today it is almost always text -- transcripts and captions retrieved by similarity. This serves the…
RLVP: Penalize the Path, Reward the Outcome
Bojie Li, Noah Shi
Agents acting on our behalf in the real world (e.g. placing phone calls) must learn online from costly, often irreversible interactions rather than cheap simulator steps. Two thing…
OneBarrier: What a Network Must Provide for Transparent Fault Tolerance to Be Free
Bojie Li
Transparent fault tolerance -- making an unmodified server binary survive crashes -- has been pursued for four decades without reaching production. Every attempt paid three costs o…
Fine-Grained Computation Offload for Off-the-Shelf Servers in Tens of Lines
Bojie Li
Hardware accelerators now sit on the critical path of online serving. GPUs, FPGAs, and increasingly remote services such as hardware security modules, post-quantum KEMs, and infere…
Whose Side Is Your Agent On? Multi-Party Principal Loyalty in LLM Agents
Bojie Li, Noah Shi
A rapidly growing class of LLM agents is multi-party: the agent acts for a principal (who briefs it, sends follow-ups, and receives results) while also conversing in a separate cha…
Agent-Computer Observation Interfaces Enable Dynamic Computer Use
Bojie Li, Noah Shi
SWE-agent established the action interface as an underexplored design axis for software-engineering agents; we make the analogous case for the observation interface in computer-use…
The Latent Bridge: A Continuous Slow-Fast Channel for Real-Time Game Agents
Bojie Li, Noah Shi
A real-time agent for general computer use - with games as the most demanding case - must act within tens of milliseconds while still planning over seconds. These two regimes sit a…
User as Engram: Internalizing Per-User Memory as Local Parametric Edits
Bojie Li
Personal memory in a language model is two problems: content and reasoning skill. The brain keeps the two apart (a sparse, local engram in the hippocampus for each episode, a slow…
PreAct: Computer-Using Agents that Get Faster on Repeated Tasks
Bojie Li
Computer-using agents drive real software through the screen -- clicking and typing -- but they solve every task from scratch: asked to repeat a task, an agent re-reads the screen,…
User as Code: Executable Memory for Personalized Agents
Bojie Li
A personalized AI agent needs a user memory: a persistent model of who the user is, built across many conversations and consulted on each new one. Today this memory is almost alway…
Models Take Notes at Prefill: KV Cache Can Be Editable and Composable
Bojie Li
Prefix caching reuses prefill only across an exactly shared prefix, so one changed field invalidates the entire downstream cache. Yet overwriting the field's own key/value vectors…
Tiara: A Programmable Line-Rate ISA for Remote Memory Access
Bojie Li
RDMA one-sided verbs are the natural primitive for memory disaggregation, but they require the client to supply the exact remote address. The 1-RTT performance breaks down when the…
OpenURMA: A Clean-Room Open Implementation of the Unified Bus Protocol
Bojie Li
Modern datacenter RDMA is bottlenecked at the network interface, not the wire. A NIC running RoCE or InfiniBand holds per-connection state for every (application, remote-endpoint)…