Karl Weinmeister
@kweinmeister.bsky.social
Cloud Developer Advocacy @ Google. AI/ML/Data, Blue Devil & Longhorn, wanna-be at home improvement. Opinions are my own.
I put antigravity.google to the test as my daily editor. Here's my story of how building a full-stack app went.
November 30, 2025 at 6:05 PM
I put antigravity.google to the test as my daily editor. Here's my story of how building a full-stack app went.
Agent Sandbox debuted at #Kubecon, unlocking a critical capability for agents: running untrusted actions. Safely validate generated code, test user input, and more.
Blog post: cloud.google.com/blog/product...
Documentation: docs.cloud.google.com/kubernetes-e...
Code: github.com/kubernetes-s...
Blog post: cloud.google.com/blog/product...
Documentation: docs.cloud.google.com/kubernetes-e...
Code: github.com/kubernetes-s...
November 11, 2025 at 9:08 PM
Agent Sandbox debuted at #Kubecon, unlocking a critical capability for agents: running untrusted actions. Safely validate generated code, test user input, and more.
Blog post: cloud.google.com/blog/product...
Documentation: docs.cloud.google.com/kubernetes-e...
Code: github.com/kubernetes-s...
Blog post: cloud.google.com/blog/product...
Documentation: docs.cloud.google.com/kubernetes-e...
Code: github.com/kubernetes-s...
Now you can see two websites in a single tab, with Chrome's new Split View.
Enable with "Split View" flag at chrome://flags.
⚠️ As an experimental feature, it could change or be removed at any time.
#browser #googlechrome #techtips
Enable with "Split View" flag at chrome://flags.
⚠️ As an experimental feature, it could change or be removed at any time.
#browser #googlechrome #techtips
November 2, 2025 at 10:22 AM
Now you can see two websites in a single tab, with Chrome's new Split View.
Enable with "Split View" flag at chrome://flags.
⚠️ As an experimental feature, it could change or be removed at any time.
#browser #googlechrome #techtips
Enable with "Split View" flag at chrome://flags.
⚠️ As an experimental feature, it could change or be removed at any time.
#browser #googlechrome #techtips
October 17, 2025 at 9:33 PM
vLLM TPU is now powered by tpu-inference!
* Broader model coverage and feature support
* 5x faster performance than Feb 2025 version
Learn more: blog.vllm.ai/2025/10/16/v...
* Broader model coverage and feature support
* 5x faster performance than Feb 2025 version
Learn more: blog.vllm.ai/2025/10/16/v...
October 16, 2025 at 8:41 PM
vLLM TPU is now powered by tpu-inference!
* Broader model coverage and feature support
* 5x faster performance than Feb 2025 version
Learn more: blog.vllm.ai/2025/10/16/v...
* Broader model coverage and feature support
* 5x faster performance than Feb 2025 version
Learn more: blog.vllm.ai/2025/10/16/v...
What API call are you making 10x a day? 🤔 Turn it into a simple /command.
With Gemini CLI extensions, you can build your own shortcuts to speed up your work:
/fetch_jira ABC-123
/deploy_staging
Learn how easy it is to get started:
geminicli.com/docs/extensi...
With Gemini CLI extensions, you can build your own shortcuts to speed up your work:
/fetch_jira ABC-123
/deploy_staging
Learn how easy it is to get started:
geminicli.com/docs/extensi...
October 8, 2025 at 3:56 PM
What API call are you making 10x a day? 🤔 Turn it into a simple /command.
With Gemini CLI extensions, you can build your own shortcuts to speed up your work:
/fetch_jira ABC-123
/deploy_staging
Learn how easy it is to get started:
geminicli.com/docs/extensi...
With Gemini CLI extensions, you can build your own shortcuts to speed up your work:
/fetch_jira ABC-123
/deploy_staging
Learn how easy it is to get started:
geminicli.com/docs/extensi...
Has Gemini ever felt like it's losing focus as your conversation goes on? Naturally, more context means more topics to cover.
Use the /compress command to keep the Gemini CLI on track. It prunes the history without a full reboot.
Get started today with the Gemini CLI: npx @google/gemini-cli
Use the /compress command to keep the Gemini CLI on track. It prunes the history without a full reboot.
Get started today with the Gemini CLI: npx @google/gemini-cli
October 7, 2025 at 5:03 PM
Has Gemini ever felt like it's losing focus as your conversation goes on? Naturally, more context means more topics to cover.
Use the /compress command to keep the Gemini CLI on track. It prunes the history without a full reboot.
Get started today with the Gemini CLI: npx @google/gemini-cli
Use the /compress command to keep the Gemini CLI on track. It prunes the history without a full reboot.
Get started today with the Gemini CLI: npx @google/gemini-cli
The Accelerate AI with Cloud Run tour is headed to Europe!
👉 Register at goo.gle/accelerate-ai
Join me and the Google team in a hands-on workshop near you:
📍 Dublin: Oct 29 goo.gle/accelerate-ai-dublin
📍 Munich: Oct 31 goo.gle/accelerate-ai-munich
📍 Paris: Nov 4 goo.gle/accelerate-ai-paris
👉 Register at goo.gle/accelerate-ai
Join me and the Google team in a hands-on workshop near you:
📍 Dublin: Oct 29 goo.gle/accelerate-ai-dublin
📍 Munich: Oct 31 goo.gle/accelerate-ai-munich
📍 Paris: Nov 4 goo.gle/accelerate-ai-paris
September 19, 2025 at 3:08 PM
The Accelerate AI with Cloud Run tour is headed to Europe!
👉 Register at goo.gle/accelerate-ai
Join me and the Google team in a hands-on workshop near you:
📍 Dublin: Oct 29 goo.gle/accelerate-ai-dublin
📍 Munich: Oct 31 goo.gle/accelerate-ai-munich
📍 Paris: Nov 4 goo.gle/accelerate-ai-paris
👉 Register at goo.gle/accelerate-ai
Join me and the Google team in a hands-on workshop near you:
📍 Dublin: Oct 29 goo.gle/accelerate-ai-dublin
📍 Munich: Oct 31 goo.gle/accelerate-ai-munich
📍 Paris: Nov 4 goo.gle/accelerate-ai-paris
The all-new Gemini 2.5 Flash Image model is built different. I had a blast mixing together multiple images and text!
👉 Try it out in AI studio: ai.dev
📖 Read all about it: developers.googleblog.com/en/introduci...
👉 Try it out in AI studio: ai.dev
📖 Read all about it: developers.googleblog.com/en/introduci...
August 26, 2025 at 2:15 PM
The all-new Gemini 2.5 Flash Image model is built different. I had a blast mixing together multiple images and text!
👉 Try it out in AI studio: ai.dev
📖 Read all about it: developers.googleblog.com/en/introduci...
👉 Try it out in AI studio: ai.dev
📖 Read all about it: developers.googleblog.com/en/introduci...
Gradio is now pre-installed in Google Colab!
It’s easier than ever to include demos and visualizations in your notebooks.
It’s easier than ever to include demos and visualizations in your notebooks.
July 23, 2025 at 12:14 PM
Gradio is now pre-installed in Google Colab!
It’s easier than ever to include demos and visualizations in your notebooks.
It’s easier than ever to include demos and visualizations in your notebooks.
Batch Mode is now supported in the Gemini API!
Submit thousands of prompts as a single job, reducing costs and simplifying your infrastructure.
✅ Priced at 50% less
✅ Higher rate limits
✅ Results in < 24 hours
👉 Read the full guide here: goo.gle/4eBuaV6
Submit thousands of prompts as a single job, reducing costs and simplifying your infrastructure.
✅ Priced at 50% less
✅ Higher rate limits
✅ Results in < 24 hours
👉 Read the full guide here: goo.gle/4eBuaV6
July 9, 2025 at 5:38 PM
Batch Mode is now supported in the Gemini API!
Submit thousands of prompts as a single job, reducing costs and simplifying your infrastructure.
✅ Priced at 50% less
✅ Higher rate limits
✅ Results in < 24 hours
👉 Read the full guide here: goo.gle/4eBuaV6
Submit thousands of prompts as a single job, reducing costs and simplifying your infrastructure.
✅ Priced at 50% less
✅ Higher rate limits
✅ Results in < 24 hours
👉 Read the full guide here: goo.gle/4eBuaV6
A deployed app is just the beginning. How do you evolve it?
You can connect Jules, an AI agentic coding assistant, to your codebase on GitHub. After requesting a new feature, Jules provides a plan for you to review. It's like having a true AI pair programmer, with the developer always in control.
You can connect Jules, an AI agentic coding assistant, to your codebase on GitHub. After requesting a new feature, Jules provides a plan for you to review. It's like having a true AI pair programmer, with the developer always in control.
June 16, 2025 at 2:07 PM
A deployed app is just the beginning. How do you evolve it?
You can connect Jules, an AI agentic coding assistant, to your codebase on GitHub. After requesting a new feature, Jules provides a plan for you to review. It's like having a true AI pair programmer, with the developer always in control.
You can connect Jules, an AI agentic coding assistant, to your codebase on GitHub. After requesting a new feature, Jules provides a plan for you to review. It's like having a true AI pair programmer, with the developer always in control.
The result: A secure, public HTTPS endpoint for your live app in seconds.
It's running on fully managed, serverless infrastructure that auto-scales from zero, ensuring cost-efficiency with zero manual configuration.
It's running on fully managed, serverless infrastructure that auto-scales from zero, ensuring cost-efficiency with zero manual configuration.
June 16, 2025 at 2:07 PM
The result: A secure, public HTTPS endpoint for your live app in seconds.
It's running on fully managed, serverless infrastructure that auto-scales from zero, ensuring cost-efficiency with zero manual configuration.
It's running on fully managed, serverless infrastructure that auto-scales from zero, ensuring cost-efficiency with zero manual configuration.
Ready to go live? By clicking "Deploy to Cloud Run," an automated CI/CD pipeline executes to store your web assets in Cloud Storage and serve them in an optimized container on Cloud Run.
June 16, 2025 at 2:07 PM
Ready to go live? By clicking "Deploy to Cloud Run," an automated CI/CD pipeline executes to store your web assets in Cloud Storage and serve them in an optimized container on Cloud Run.
It all starts in Google AI Studio. Instead of writing boilerplate, you describe your application's purpose in natural language.
Gemini generates a complete, well-structured project with a UI and application logic, giving you an interactive preview instantly.
Gemini generates a complete, well-structured project with a UI and application logic, giving you an interactive preview instantly.
June 16, 2025 at 2:07 PM
It all starts in Google AI Studio. Instead of writing boilerplate, you describe your application's purpose in natural language.
Gemini generates a complete, well-structured project with a UI and application logic, giving you an interactive preview instantly.
Gemini generates a complete, well-structured project with a UI and application logic, giving you an interactive preview instantly.
"Vibe coding" is fun, but what do you do after the code is generated?
"Vibe deployment" can take you from a prompt to a deployed, scalable app in seconds. Here's a look at the entire process.
"Vibe deployment" can take you from a prompt to a deployed, scalable app in seconds. Here's a look at the entire process.
June 16, 2025 at 2:07 PM
"Vibe coding" is fun, but what do you do after the code is generated?
"Vibe deployment" can take you from a prompt to a deployed, scalable app in seconds. Here's a look at the entire process.
"Vibe deployment" can take you from a prompt to a deployed, scalable app in seconds. Here's a look at the entire process.
How could you discover clues that content might be AI-generated?
I found top words and key insights from #Gemini Flash 2.5:
🧭 Roadmapping the answer to come: "Let's breakdown"
🧐 Definitive answers: "crucial" and "comprehensive"
🙌 Enthusiastic language: "great question"
I found top words and key insights from #Gemini Flash 2.5:
🧭 Roadmapping the answer to come: "Let's breakdown"
🧐 Definitive answers: "crucial" and "comprehensive"
🙌 Enthusiastic language: "great question"
June 5, 2025 at 3:46 PM
How could you discover clues that content might be AI-generated?
I found top words and key insights from #Gemini Flash 2.5:
🧭 Roadmapping the answer to come: "Let's breakdown"
🧐 Definitive answers: "crucial" and "comprehensive"
🙌 Enthusiastic language: "great question"
I found top words and key insights from #Gemini Flash 2.5:
🧭 Roadmapping the answer to come: "Let's breakdown"
🧐 Definitive answers: "crucial" and "comprehensive"
🙌 Enthusiastic language: "great question"
This week, the Agent2Agent Python SDK has been released, along with an updated v0.2 spec.
To showcase these new capabilities, the Agentic Trading sample has been updated. A trader agent proposes trades and a risk agent approves these trades.
👉 github.com/kweinmeister...
To showcase these new capabilities, the Agentic Trading sample has been updated. A trader agent proposes trades and a risk agent approves these trades.
👉 github.com/kweinmeister...
May 22, 2025 at 8:37 PM
This week, the Agent2Agent Python SDK has been released, along with an updated v0.2 spec.
To showcase these new capabilities, the Agentic Trading sample has been updated. A trader agent proposes trades and a risk agent approves these trades.
👉 github.com/kweinmeister...
To showcase these new capabilities, the Agentic Trading sample has been updated. A trader agent proposes trades and a risk agent approves these trades.
👉 github.com/kweinmeister...
Full codebase understanding! 🧠
May 14, 2025 at 9:03 PM
Full codebase understanding! 🧠
Yek prioritizes files within a token budget, using Git history & other rules. github.com/bodo-run/yek
Command: yek --tokens 1024k .
Result: ~253k tokens! 🎉 Room to spare.
Command: yek --tokens 1024k .
Result: ~253k tokens! 🎉 Room to spare.
May 14, 2025 at 9:03 PM
Yek prioritizes files within a token budget, using Git history & other rules. github.com/bodo-run/yek
Command: yek --tokens 1024k .
Result: ~253k tokens! 🎉 Room to spare.
Command: yek --tokens 1024k .
Result: ~253k tokens! 🎉 Room to spare.
Still using Repomix, added the --compress flag to strip comments & whitespace from the filtered set.
Command: repomix . --ignore "..." --compress
Result: Down to ~1.8M tokens. Getting closer!
Command: repomix . --ignore "..." --compress
Result: Down to ~1.8M tokens. Getting closer!
May 14, 2025 at 9:03 PM
Still using Repomix, added the --compress flag to strip comments & whitespace from the filtered set.
Command: repomix . --ignore "..." --compress
Result: Down to ~1.8M tokens. Getting closer!
Command: repomix . --ignore "..." --compress
Result: Down to ~1.8M tokens. Getting closer!
Used Repomix again, this time to filter out non-essential file types for LLM understanding (data files, static assets, etc.).
Command: repomix . --ignore "**/*.ipynb,**/*.csv,..."
Result: Down to ~2.6M tokens! Massive cut.
Command: repomix . --ignore "**/*.ipynb,**/*.csv,..."
Result: Down to ~2.6M tokens! Massive cut.
May 14, 2025 at 9:03 PM
Used Repomix again, this time to filter out non-essential file types for LLM understanding (data files, static assets, etc.).
Command: repomix . --ignore "**/*.ipynb,**/*.csv,..."
Result: Down to ~2.6M tokens! Massive cut.
Command: repomix . --ignore "**/*.ipynb,**/*.csv,..."
Result: Down to ~2.6M tokens! Massive cut.
Used another tool for a second opinion on size.
Tool: gitingest.com
Command: gitingest .
Result: ~69 MILLION tokens. Confirmed: It's HUGE.
Tool: gitingest.com
Command: gitingest .
Result: ~69 MILLION tokens. Confirmed: It's HUGE.
May 14, 2025 at 9:03 PM
Used another tool for a second opinion on size.
Tool: gitingest.com
Command: gitingest .
Result: ~69 MILLION tokens. Confirmed: It's HUGE.
Tool: gitingest.com
Command: gitingest .
Result: ~69 MILLION tokens. Confirmed: It's HUGE.
First, got a baseline token count for the python-docs-samples repo.
Tool: github.com/yamadashy/re... from
@yamadashy.bsky.social
Command: repomix .
Result: ~56 MILLION tokens! 🤯
Tool: github.com/yamadashy/re... from
@yamadashy.bsky.social
Command: repomix .
Result: ~56 MILLION tokens! 🤯
May 14, 2025 at 9:03 PM
First, got a baseline token count for the python-docs-samples repo.
Tool: github.com/yamadashy/re... from
@yamadashy.bsky.social
Command: repomix .
Result: ~56 MILLION tokens! 🤯
Tool: github.com/yamadashy/re... from
@yamadashy.bsky.social
Command: repomix .
Result: ~56 MILLION tokens! 🤯
🚀 Qwen3 has dropped!
📐 0.6B - 235B model sizes: “Qwen3-4B can rival the performance of Qwen2.5-72B-Instruct”
🧠 Hybrid thinking and non-thinking modes
🤖 Improved agentic capabilities
🌎 Multilingual support with 119 languages
Blog: qwenlm.github.io/blog/qwen3/
It’s working great on Google Cloud!
📐 0.6B - 235B model sizes: “Qwen3-4B can rival the performance of Qwen2.5-72B-Instruct”
🧠 Hybrid thinking and non-thinking modes
🤖 Improved agentic capabilities
🌎 Multilingual support with 119 languages
Blog: qwenlm.github.io/blog/qwen3/
It’s working great on Google Cloud!
April 29, 2025 at 2:25 AM
🚀 Qwen3 has dropped!
📐 0.6B - 235B model sizes: “Qwen3-4B can rival the performance of Qwen2.5-72B-Instruct”
🧠 Hybrid thinking and non-thinking modes
🤖 Improved agentic capabilities
🌎 Multilingual support with 119 languages
Blog: qwenlm.github.io/blog/qwen3/
It’s working great on Google Cloud!
📐 0.6B - 235B model sizes: “Qwen3-4B can rival the performance of Qwen2.5-72B-Instruct”
🧠 Hybrid thinking and non-thinking modes
🤖 Improved agentic capabilities
🌎 Multilingual support with 119 languages
Blog: qwenlm.github.io/blog/qwen3/
It’s working great on Google Cloud!