Douglas Lai

Founding Member | Product Growth & Design @Eigent_AI & @CamelAIOrg . Prev @ImperialDyson @RCA @PrattInstitute Keio

[Douglas Lai @douglas_ym]

Built this GLM-5.2 comparison use case on Eigent for long horizon task: deeper planning, wider research, and stronger AI infra research → JSON → HTML report performance.

[Eigent AI @Eigent_AI]

@Zai_org GLM 5.2 is now live on Eigent from day 0!

threw a real long-horizon task: research 30 companies across 6 sectors of the AI infrastructure stack, structure it into JSON, then build an interactive HTML report. same prompt, 5.1 vs 5.2. where 5.2 pulls ahead:

→ plans deeper and verifies harder.
→ researches wider. Covers 2x the ground.
→ ships a more complete, interactive report.

with a 1M context window, it holds the whole task from research to final deliverable. no dropped threads, no half-finished output.

long tasks are where the difference shows. try it on open source coworker Eigent. 🦾

[Douglas Lai @douglas_ym]

Open source models are getting seriously good!
I made this use case with @deepseek_ai V4 via @ollama , and now use it weekly/monthly for our product development reports — PRs → doc → Slack update, all in one prompt with @Eigent_AI .

[Eigent AI @Eigent_AI]

@ollama + @deepseek_ai v4 pro handled entire monthly dev reports on Eigent. github prs → word doc → slack message → sent to product-release channel. in just one prompt. fully local.

the full walkthrough is in the thread. try the same loop on open source cowork Eigent, byo model!

[Douglas Lai @douglas_ym]

@trondw @Eigent_AI @Google 🥳🤝 yesss, excited for more Gemini collabs ahead!

[Douglas Lai @douglas_ym]

A small surreal moment.

Drinking my favourite beer, looking at the logo I designed, and seeing @Eigent_AI, the product we’re building —— show up at @Google I/O 2026.

#GoogleIO #AIProduct #Eigent

[Douglas Lai @douglas_ym]

This is the kind of agent handoff workflow that makes multi-agent systems feel practical.

A real Megatron-LM CI failure went from messy logs to root cause in minutes with Gemini Managed Agents on Eigent.

[Eigent AI @Eigent_AI]

Gemini 3.5 flash + Gemini managed agents api just audited a real megatron-lm ci failure inside Eigent. root cause in minutes!

watch the handoff: coordinator agent plans the audit, developer agent loads the ml-failure-audit skill and gathers the evidence, then gemini agent steps in as a remote sub-agent for the heavy reasoning.

gemini managed agents api and gemini 3.5 flash now live on our open-source cowork Eigent! @googleaidevs
@GoogleAIStudio

[Eigent AI @Eigent_AI]

Gemini 3.5 flash now live on Eigent! ⚡️
You can byok or select it as a cloud model in our open source cowork. a use case is cooking. drops soon 👀

[Douglas Lai @douglas_ym]

Single agent or multi-agent — which is better? 🧐

[Douglas Lai @douglas_ym]

Cooking up something smooth and crafted with micro-interactions for a brand-new version of Eigent.

https://t.co/69gFIBFFtG

[Douglas Lai @douglas_ym]

People testing Claude Design with one-sentence benchmarks are measuring the wrong thing. Good design cannot live without context. The value is in making it easy to connect code, files, and assets to build design systems and the infrastructure for better design work.

[Claude @claudeai]

Introducing Claude Design by Anthropic Labs: make prototypes, slides, and one-pagers by talking to Claude.

Powered by Claude Opus 4.7, our most capable vision model. Available in research preview on the Pro, Max, Team, and Enterprise plans, rolling out throughout the day.

[Douglas Lai @douglas_ym]

Testing Claude Design with one-sentence tasks is nonsense. Great design does not exist without context.

What makes Claude Design interesting is not another design agent, but how easily it connects code, files, and assets to build a real design system and a design engine.

[Douglas Lai @douglas_ym]

Thanks @github for the invite — representing @CAMELAIOrg co-hosting GitHub Social Club #London. Great to meet so many devs, builders, maintainers, and founders. Lovely vibe, great conversations, and a strong open-source community. #CAMELAI #GitHubSocialClub #OpenSource

[Douglas Lai @douglas_ym]

BAZI.md is a Chinese-culture-inspired agent standard for defining an AI agent’s underlying makeup:
- temperament
- decision bias
- collaboration posture
- risk boundaries
- future tendency

Repo: https://t.co/rCEC67cc0G
Demo: https://t.co/yrn24Uhiw6

[Douglas Lai @douglas_ym]

AGENTS.md, SKILLS.md, DESIGN.md…

We’ve been defining more and more of AI agents in markdown. But most of these specs describe only instructions.

BAZI.md explores a different idea: What defines an agent’s underlying nature? Not instructions. Identity.

https://t.co/yrn24Uhiw6

[Douglas Lai @douglas_ym]

I’ll be co-hosting @github GitHub Social Club in #London on April 7, representing @CamelAIOrg . If you’re into AI, agentic products, open source, or just want to meet people building interesting things, come hang out.

https://t.co/4Ed9HC1J6b

[Douglas Lai @douglas_ym]

@Eigent_AI Love it! open-source AI coworkers are catching up to Claude’s cowork features fast 👀

[Eigent AI @Eigent_AI]

Eigent × @Kimi_Moonshot K2.5

We ran a real-world sales performance evaluation on Eigent using Kimi K2.5.

Given a production sales dataset, Eigent coordinated multiple agents end to end:
- The document agent extracts information from the source files
- The terminal agent analyzes the data and generates an HTML report

All results are stored locally! Try Eigent with Kimi K2.5 today.

[Douglas Lai @douglas_ym]

@wabi build me an app that lets me photograph crystals from my collection and save each one as an item in a digital crystal library. For every saved crystal, generate a dedicated “crystal space” page I can open anytime for a meditation session

[Douglas Lai @douglas_ym]

@wabi build me an app that lets me photograph crystals from my collection and save each one as an item in a digital crystal library. For every saved crystal, generate a dedicated and personalised “crystal space” page I can open anytime for a meditation session.

[Guohao Li 🐫 @guohao_li]

Running Eigent locally with @lmstudio!
https://t.co/gCHQz83Rrr

[Guohao Li 🐫 @guohao_li]

The SEA Workshop at @NeurIPSConf 2025 is coming next Sunday. It seems we urgently need more open, realistic agent environments for training and evaluating agents. But what are the important environments to build? What are the infrastructure bottlenecks for these environments in training and evaluation? How can we scale up the number of available environments? And how should we use these environments, RL or beyond? These questions are still not clear.

We’re bringing together an amazing list of speakers and panelists to spark the discussion: @egrefen, @Mike_A_Merrill, @mialon_gregoire, @deepaknathani11, @jl_marino, @syz0x1, @qhwang3, Anthony G. Cohn, Eric Sommerlade, and @fredsala. You won’t want to miss it if you’re around.

Also, huge thanks to our four sponsors, @TheInclusionAI (@AntLingAGI), @SnorkelAI, @SonicjobsApp, and @VmaxAI for their generous support!