Wai Hon Law (@whhone) profile picture

Software Engineer, Emacs User, Runner

Joined December 2009
Photo by Wai Hon Law (@whhone)Photo by Wai Hon Law (@whhone)Photo by Wai Hon Law (@whhone)Photo by Wai Hon Law (@whhone)Photo by Wai Hon Law (@whhone)Photo by Wai Hon Law (@whhone)Photo by Wai Hon Law (@whhone)Photo by Wai Hon Law (@whhone)
Introducing Gemini 4 Argon – our new frontier model. It’s built for complex workflows across coding, enterprise knowledge work, and cybersecurity defense – rolling out today to a set of trusted testers through our Fairwind Program.
1,986
4,629
4,527
42,776
9,037,670
Run open models like Gemma 4 completely offline in the Antigravity SDK. Built on Google AI Edge’s LiteRT, you can now run Gemma 4 directly on your local GPU. Zero API costs, total data privacy, and no internet required.
174
565
117
5,657
4,568,593
Antigravity SDK August 31st 2026 weekly update for v0.1.16: - 🎉 Gemini 3.8 Flash Launch: The SDK now defaults to gemini-3.8-flash landing large improvements to coding capabilities and overall reasoning! - Announcement: blog.google/innovation-and-a… - 🎉 Terminal Sandboxing: Run shell commands in an OS-level sandbox by setting enable_sandbox on RunCommandConfig. - Execution Stop Hook: Register a StopHook to intercept turn cancellations, execute cleanup handlers, and persist state before an agent halts. - Lightweight Agent Preset: Call AgentConfig.lightweight() to configure minimal tools and token compaction limits to optimize a lower token heavy Antigravity experience. - Callable Class Tools: Pass callable class instances and functors directly into the tools list with automatic parameter and context resolution. github.com/google-antigravit… 🧵
3
3
58
4,433
Wai Hon Law retweeted
Excited to launch Gemma 4: the best open models in the world for their respective sizes. Available in 4 sizes that can be fine-tuned for your specific task: 31B dense for great raw performance, 26B MoE for low latency, and effective 2B & 4B for edge device use - happy building!
319
850
178
7,887
998,841
Mobile Actions and Tiny Garden are available out-of-the-box in the Google AI Edge Gallery app, now on both Android and iOS. Benchmark the model performance directly in the app. Discover how you can use FunctionGemma (270M) and LiteRT-LM to build private, high-performance local assistants: goo.gle/3ODJJCT
16
67
13
605
59,171
Wai Hon Law retweeted
New course: Gemini CLI: Code & Create with an Open-Source Agent, built with @googlecloudtech/@geminicli and taught by @JackWoth98. Agentic coding assistants like Gemini CLI are transforming how developers work. This short course teaches you to use Google's open-source agent to coordinate local tools and cloud services for coding and non-coding workflows. Gemini CLI works from your terminal, so it works with your local files and development tools. You can also connect it to services through MCP. Then provide high-level instructions, and it autonomously plans and executes complex workflows. Skills you'll gain: - Build website features and automate code reviews with GitHub ActionsCreate data dashboards that combine local files with cloud data sources - Use MCP servers and extensions to orchestrate workflows across GitHub, Canva, and Google Workspace - Generate social media content from multimedia files like conference recordings I particularly appreciate that Gemini CLI is open-source. You can see exactly how it works, read the prompts it uses, and understand its architecture. The community has contributed thousands of pull requests. Since Gemini 3’s release I've found Gemini CLI highly capable - this is a tool worth having in your toolbox! Whether you're prototyping applications, automating workflows, or working with multimedia content, join to learn to delegate complex tasks and build faster: deeplearning.ai/short-course…
52
188
11
1,327
126,280
Introducing T5Gemma 2, the next generation of encoder-decoder models 🚀 Built on top of Gemma 3, we were able to build compact models at sizes of 270m-270m, 1B-1B, and 4B-4B sizes. While most models today are decoder-only, T5Gemma 2 is the first (I'm aware of) multimodal, long-context, and heavily multilingual (140 languages) encoder-decoder model out there. We hope this model enables the model research community as well as the community of devs ready to explore with new architectures. Blog: blog.google/technology/devel… Models: huggingface.co/collections/g… Paper: arxiv.org/abs/2512.14856
75
165
28
1,448
248,430
We're blown away by the great discussions and feedback shared, and we’re actively working on improvements. A new Gallery app update is out now with fixes based on your input.
1
8
1
22
9,347
Wai Hon Law retweeted
Introducing Umbrel Home – a home server that puts your data back in your hands. Run your own: - Cloud storage - Bitcoin node - Media/VPN/home automation server - Network adblocker - 100s of self-hosted apps Home is where your data belongs. Pre-order: umbrel.com
138
257
120
1,308
368,660
工程师成长到最后最重要的是什么? chen.cat/3my3XzL
3
15
6
96
59,526
Pushing The "Hardcore" Button. There are times in my career when I would have already pushed the Hardcore Button. My self-image was based on out-working, out-intensitying everyone. I wouldn't push it now, but I can understand if someone else does. Here are 3 strategies.
23
113
37
651
And the AI madness continues with the Google Pixel 7 Pro. 1st Photo - 1x 2nd Photo - 30x
94
280
41
6,015
Shouting at disk will increase its latency... youtu.be/tDacjrSCeq4 via @YouTube
面试题:开发新功能和重构老代码之间怎么选? chen.cat/3dubdYW
7
9
111
美国互联网公司大多数都是要撤出中国的 chen.cat/3RdwxRZ
5
1
27
Wai Hon Law retweeted
Replying to @BenBear
Most OKR exercises are just random ass goals tacked on to existing priorities and roadmaps, and almost always miss the Why. This format has helped with my teams in the past. Also, generally the key results aren’t time and measurement bound enough at first.
5
25
4
380