The job market and AI First
Why AI first when working on or creating new workflows? Mostly because its in the interest of your career to show measurable wins in the AI space that you can fold into your own CV. As this article highlights a good starting point is bullet points showing specific tool and model used, real task performed, fiction solved, outcome and the judgement shown.
https://substack.com/home/post/p-161426515
AI is rapidly surpassing humans in coding, with OpenAI, Anthropic, Google, and Meta releasing powerful models like GPT-4.1 and Claude Code, dramatically accelerating software development and transforming developer roles.
“This is the year that AI becomes better than humans at competitive code forever,” said OpenAI’s chief product officer Kevin Weil. He compared the advances to AI surpassing humans at chess, but argued this had a more democratising impact “on the world if everybody can create software”.
https://www.ft.com/content/8069b127-8589-4f06-9c38-8e0216c6fd9c
Claude Code Origin Story & Demo Webinar
I am not going to lie I was pretty blown away by this, I feel like I am living in the future.
Video here
Presentation deck
Facts:
I am spending the Easter weekend completing this CrewAI agent course. Its amazing how easy it is to get immediately hands on using Google Collab and LLM API keys to setup some very cool applications. I imagine very quickly a lot of POCs of new features and prototypes will be completed by Product Managers in this way and then pushed rapidly to PRD with some validation steps by engineering
https://www.deeplearning.ai/short-courses/multi-ai-agent-systems-with-crewai/
At TED 2025, OpenAI CEO Sam Altman discussed the company's rapid growth to 800 million weekly users, infrastructure challenges around AI parlour tricks like turning dogs into humans or Studio Ghibli images , artist compensation plans, and ethical concerns surrounding autonomous AI agents
Speaking of AI parlour tricks here is my dogs group of friends as humans, although it misgendered my dog Charlie (white cockapoo at front of image)
Gemini now has advanced Veo Video creation
https://www.t3.com/tech/google-just-added-new-gemini-video-creation-tool
Sadly not appearing for me
This article outlines a structured approach to "vibe coding" full-stack apps using AI tools like Cursor and Gemini 2.5 Pro, emphasising strong foundations, clear rules, and ongoing documentation for efficient development.
https://dev.to/wasp/a-structured-workflow-for-vibe-coding-full-stack-apps-352l
He is playing Chess while you play Chequers
Google has published a 68 page whitepaper focused on Prompt Engineering (focused on API users), and it goes deep on structure, formatting, config settings, and real examples.Here’s what it covers:
How to get predictable, reliable output using temperature, top-p, and top-k
Prompting techniques for APIs, including system prompts, chain-of-thought, and ReAct (i.e., reason and act)
How to write prompts that return structured outputs like JSON or specific formats
PDF here: Prompt Engineering Whitepaper (Google, 2025)
To be fair Google images has been able to find location from images for years so not sure what is new here…..
Microsoft going big on the Orwellian features with Recall
https://www.bbc.co.uk/news/articles/cj3xjrj7v78o.amp
Anthropic Claude has released a Deep Research feature and its getting good reviews and comparable with other foundation offerings.
I do not currently use NotebookLM but perhaps its time I started
https://www.xda-developers.com/finally-started-using-notebooklm-and-should-have-sooner/
True Story
