Y Combinator just dropped the vibe coding playbook

Tom Blomfield (ex-Monzo, now YC) has been testing the edge of AI-assisted dev. After a month of building side projects with Claude, Windsurf, and Aqua, he’s convinced: LLMs aren’t just copilots — they’re almost full-stack collaborators.

My take:

• “Vibe coding” isn’t hype. It’s the emerging skill of rapid, AI-led prototyping — blending intuition, prompts, and live iteration.

• YC’s approach is disciplined: version control, tests, documentation, refactoring — plus leaning into LLMs for debugging, scaffolding, and even planning.

• This isn’t no-code. It’s augmented code — where a sharp human with clear intent can ship working products 5–10x faster.

• One killer tip: treat LLMs like a sounding board, not just a code generator. Ask it to challenge your stack choices, explain trade-offs, or write tests.

• “Keep experimenting” isn’t an afterthought — it’s the muscle that separates prompters from builders.

Key part: “AI can’t one-shot a whole product… yet. But it can absolutely help you vibe your way to MVP velocity.”

Watch the video on YouTube

Google just dropped 10 completely free AI courses

Google just released a new set of 10 AI courses — all free, all beginner-friendly. Whether you want to understand the basics or explore responsible AI, there’s a course for you. No technical background required. Here is the full list:

1. Introduction to Generative AI: What Generative AI is, how it is used, and how it differs from traditional machine learning methods.

2. Introduction to Large Language Models (LLMs): what large language models (LLM) are, the use cases where they can be utilised, and how you can use prompt tuning to enhance LLM performance.

3. Introduction to Responsible AI: What responsible AI is, why it's important, and how Google implements responsible AI in their products. It also introduces Google's 7 AI principles.

4. Gen AI: Unlock Foundational Concepts: In this course, you unlock the foundational concepts of generative AI by exploring the differences between AI, ML, and gen AI, and understanding how various data types enable generative AI to address business challenges. You also gain insights into Google Cloud strategies to address the limitations of foundation models and the key challenges for responsible and secure AI development and deployment.

5. Introduction to Image Generation: This course introduces diffusion models, a family of machine learning models that recently showed promise in the image generation space.

6. Encoder Decoder Architecture: A synopsis of the encoder-decoder architecture, which is a powerful and prevalent machine learning architecture for sequence-to-sequence tasks such as machine translation, text summarisation, and question answering.

7. Attention Mechanisms: This course will introduce you to the attention mechanism, a powerful technique that allows neural networks to focus on specific parts of an input sequence.

8. Transformers and BERT models: Transformer architecture and the Bidirectional Encoder Representations from Transformers (BERT) model.

9. Create Image Captioning Models: How to create an image captioning model by using deep learning.

10. Introduction to Vertex AI Studio: A tool to interact with generative AI models, prototype business ideas, and launch them into production.

My take:

• Google isn’t just giving away education - they’re shaping the AI talent pipeline.

• These courses aren’t fluff. They reflect where big tech wants the workforce to go.

• If you’re in consulting, product, or innovation - taking one could help you speak the language of AI with confidence.

• Don’t overthink which one to start with. Pick one, commit to an hour, and you’ll build momentum.

Key insight: Free training from tech giants is a signal - they’re betting on scale, and they want you in their ecosystem.

And don’t you want to earn Google’s new AI Badges?! 🪪
Thanks to Jimmy Acton, CDO at 4th Utility for sharing.

Duolingo is going AI-first

This week, Duolingo’s CEO announced the company is officially going AI-first. In a follow-up Slack shared by their Chief Engineering Officer, Natalie Glance, we got a rare behind-the-scenes look at how they’re actually making it happen.

My take:

• This isn’t just lip service. Duolingo’s internal guidance is sharp — it nails the real levers: abstraction, speed, and system-level change.

• “Start with AI for every task” is the killer line. It’s not about whether AI is better yet — it’s about building the habit of leverage.

• Productivity will rise — not by adding hours, but by removing bottlenecks. Faster prototypes, better docs, tighter feedback loops.

• The boldest point? 10% of time should go to learning AI. That’s a company investing in long-term capability, not just short-term output.

• Technical excellence still matters. In fact, AI raises the bar — because you can’t hide behind boilerplate when machines are helping write it.

• This is how culture shifts happen: clear playbooks, leadership consistency, and real examples from the top.

Key insight: “Being AI-first doesn’t lower the bar. It raises it.” — Natalie Glance, Duolingo

See the post on LinkedIn

Meet Zapier's AI orchestration platform

Zapier now lets you embed AI agents directly into workflows across nearly 8,000 apps. It’s a major move — and a signal that the future of AI isn’t chatbots, it’s systems.

My take:

• Zapier’s not just helping you automate tasks — it’s helping you build full AI-powered systems.

• AI agents can now decide, search, and adapt mid-Zap. This is AI meeting structured automation.

• Vendasta, Remote.com and Okta are using this to reclaim thousands of hours and millions in value.

• Their “human-in-the-loop” feature via Slack approvals is smart — it keeps teams in control while scaling.

• This platform shift is a masterclass in connecting strategy to execution. No AI silos. No six-month pilots. Just building.

If AI isn’t connected to your workflows, it’s just theatre. Zapier is quietly making it real.

Read about the launch

(also checkout Linear’s MCP Server example with Claude. Mega fun stuff!)