Sabitlenmiş Tweet
Petko D. Petkov
9.7K posts
Petko D. Petkov
@pdp
on a break from CISO duties, building https://t.co/Tdjrt4MFFN
London Katılım Şubat 2008
304 Takip Edilen5.7K Takipçiler
Most work is not content.
We trained AI to write. But most of what people do at work is not writing.
Until we look past content generation, we will never figure out how to use agents at scale.
chatbotkit.com/reflections/mo…
English
Lock down your MCP servers so only your team can use them.
Set up Google OAuth, create dedicated skillsets, and connect everything in the Blueprint Designer - step by step.
chatbotkit.com/tutorials/how-…
#MCP #OAuth #AI #developers #ChatBotKit
English
Deploy your agent army.
Build and orchestrate fleets of AI agents that work together, stay in sync, and get things done.
APIs, SDKs, batteries-included backend.
chatbotkit.com
English
Give your AI chatbot a real memory.
Learn how to build bots that remember context across conversations using ChatBotKit - no more starting from scratch.
chatbotkit.com/tutorials/how-…
#AI #chatbot #ChatBotKit
English
Three new MiMo-V2 models are now live on ChatBotKit.
Pro - 1M context window. Omni - multimodal. Flash - cost-efficient reasoning.
More models, more choices.
chatbotkit.com/changelog/intr…
#AI #agents #multimodal
English
AI isn't killing open source with one blow - it's eroding it from multiple directions at once.
Code gen reduces contributions. Model training consumes open data. Closed AI stacks replace open toolchains.
The community needs to adapt fast.
chatbotkit.com/reflections/ai…
English
Schedulers tell agents when to run. Supervisors make sure they run right.
The next leap in agentic AI is not better scheduling - it is autonomous verification.
chatbotkit.com/reflections/ag…
#AI #agents #agentic #CBK
English
GPT-5.4 Mini and GPT-5.4 Nano are now on ChatBotKit.
Smaller footprint. Serious performance. Built for coding, subagents, and high-volume workflows.
chatbotkit.com/changelog/intr…
English
API abilities as shell commands.
Install the API skillset, then pipe, loop, and script across GitHub, Slack, Linear, and HubSpot from bash - with ChatBotKit handling all the auth.
chatbotkit.com/blueprints/she…
English
Your AI agents are only as good as your last test.
The Continuous Agent Evaluation Harness stress-tests your agents automatically - catching regressions before your users ever notice.
chatbotkit.com/blueprints/con…
English
bro, it's so over for designers
google stitch is insane. 🤯
Stitch by Google@stitchbygoogle
Meet the new Stitch, your vibe design partner. Here are 5 major upgrades to help you create, iterate and collaborate: 🎨 AI-Native Canvas 🧠 Smarter Design Agent 🎙️ Voice ⚡️ Instant Prototypes 📐 Design Systems and DESIGN.md Rolling out now. Details and product walkthrough video in 🧵
English
API abilities as shell commands.
Install the API skillset, then pipe, loop, and script across GitHub, Slack, Linear, and HubSpot from bash - with ChatBotKit handling all the auth.
chatbotkit.com/blueprints/she…
English
One agent, multiple APIs, zero credential juggling.
Managed secrets + raw HTTP abilities = clean, composable API access for your agents.
chatbotkit.com/blueprints/adh…
English
Your AI agents are only as good as your last test.
The Continuous Agent Evaluation Harness stress-tests your agents automatically - catching regressions before your users ever notice.
chatbotkit.com/blueprints/con…
English
Your AI agents are only as good as your last test.
The Continuous Agent Evaluation Harness stress-tests your agents automatically - adversarial inputs, LLM-as-judge scoring, and regression alerts before your users notice.
chatbotkit.com/blueprints/con…
English
1.4K pull requests raised since January 1st, entirely by the CBK.AI agentic coding system.
No manual commits. No human PRs. Just autonomous agents shipping code at scale.
English
AI gave us unprecedented productivity. Almost nobody has figured out how to turn that productivity into products that matter.
Speed is not a strategy. Building is.
Read the full reflection: chatbotkit.com/reflections/th…
English
GPT-5.3-Codex is now on CBK.AI.
OpenAI's specialized coding model is built for autonomous software engineering - long-running code tasks, complex execution, and deep reasoning.
chatbotkit.com/changelog/intr…
English