Poolside (@poolsideai) / X

Poolside

205 posts

Poolside

@poolsideai

We build models for agentic coding and long-horizon tasks. Try Laguna: poolside.ai/get-started

San Francisco, CA

discord.gg/poolsideai

Joined May 2023

Following

13.8K

Followers

Affiliates Replies Media

Pinned
Poolside
@poolsideai
Jul 28
Today, we’re releasing Poolside Desktop Assistant. One place to run coding agents across macOS, VS Code, and Visual Studio. We built it for ourselves and have used it every day for the past year. Now we’re opening it up to everyone.
00:00
Poolside
@poolsideai
14h
Building AI in the open makes the world more secure. We’re joining the Open Secure AI Alliance to build open tools that safeguard software and agents. We’ll keep releasing model weights and evaluations and sharing our research to strengthen a broader open ecosystem for
Jensen Huang
@JensenHuang
Jul 27
Attackers have frontier AI. Defenders need a frontier AI ecosystem—the best open and closed models, force-multiplied by a global community. During the Hugging Face incident, closed AI blocked essential forensics. An open-weight frontier model helped contain the intrusion.
Poolside
@poolsideai
17h
The path to an evaluation score matters as much as the score itself. As models become more capable and persistent, reward hacking becomes a more prominent challenge for agent evaluation. It is not enough to know that an agent reached the right result. We need to understand how
Appen Research
@AppenResearch
17h
An agent can pass a benchmark and still fail the test. For Laguna S 2.1, @poolsideai evaluated full trajectories, not just final scores. @AppenResearch worked with Poolside on aspects of the reward hacking detection used during the evaluation process.
Poolside
@poolsideai
Jul 28
We completely agree. We think the interface should be independent of the model or harness. That’s exactly why we built Poolside Desktop Assistant. Download below 👇
@jason
@Jason
Jul 28
Open source, open weight models surging… now we just need app layer/harness/interfaces that make them shine Power move for @claudeai would be to allow you to use any model with Cowork — @DarioAmodei you up for that?
Poolside
@poolsideai
Jul 28
Laguna XS 2.1 already runs on a Mac. Now it's time to make it fly. We teamed up with @eigenlabs on MLX.fast - an open autoresearch competition to optimize Laguna XS 2.1 inference on Apple Silicon. Point your agent at it and improve the open weight ecosystem
The Laguna MLX Challenge
From mlx.fast
See @poolsideai's full profile
Sign up
Log in

See @poolsideai's full profile

See @poolsideai's full profile