Google Launches AI Everything, Claude Claps Back and OpenAI Goes Full Hardware

Google floods the zone, Anthropic tops key benchmarks and OpenAI teams up with Apple’s design legend for what’s next.

Fellow Data Tinkerers!

Today I’m announcing a brand new addition to my Substack publication: Data Tinkerer subscriber chat. This is a conversation space for subscribers. kind of like a group chat or live hangout. I’ll post questions and updates that come my way and you can jump into the discussions and or make suggestions.

How to get started

1. Get the Substack app by clicking this link or the button below. New chat threads won’t be sent sent via email, so turn on push notifications so you don’t miss conversation as it happens. You can also access chat on the web.

2. Open the app and tap the Chat icon. It looks like two bubbles in the bottom bar, and you’ll see a row for my chat inside.

3. That’s it! Jump into my thread to say hi, and if you have any issues, check out Substack’s FAQ.

Feel free to share your opinion or suggestions there, I’d really appreciate it.

Now let’s get to this week’s round-up!

---

The Buzz 🐝

Talk about a gangbuster week in AI with all the announcements! This meme has never been more appropriate:

All right, let’s recap the 3 main events last week:

• The highest number of announcements came out of Google I/O 2025. Rapid-fire summary:
  1- Gemini 2.5 Pro is getting a ‘deep think’ feature
  2- Flash 2.5 is improved across key benchmarks for reasoning and code
  3- Jules is the new coding agent (similar to OpenAI’s Codex)
  4- Veo 3 can now generate videos with audio. Check the example below

  5- Flow, AI filmmaking tool powered by Veo 3. Check this non-existent car show below that someone created with Flow (Source). Pretty damn good!

  And many more announcements. You can check all of them here or stuff you can build with Google AI for developers here. So Google was the leaderboard on most of the benchmarks until …

• Anthropic announced Claude 4 last week. Opus 4 now beats OpenAI’s Codex in software engineering benchmark and Sonnet 4 is the best performing model for everyday tasks.

  

  (Source: Anthropic)

  They also released a practical video of how to use Claude in your workflow:

• And the last but not the least which you might have already seen is OpenAI buying Jony Ive’s (Apple’s former Chief Design Officer) company. The entity (called io) will be merged with OpenAI and it will be focusing on creating the first hardware device for OpenAI, possibly as soon as 2027.

When your startup pivot turns into a 90s boy band reunion shoot

---

Data Science & AI

• Getting AI to write good SQL: Text-to-SQL techniques explained
  Learn how Google uses large LLMs and smart prompt engineering to make natural language to SQL translation more accurate and actually useful in real-world databases.

  

  (Source: Google)

• How Reddit Scans 1M+ Images a Day to Flag NSFW Content Using Deep Learning
  Reddit needed to flag NSFW images the second they were uploaded. They built a deep learning system that does exactly that; fast, scalable and battle-tested in prod. Here’s how it works.

• How DoorDash leverages LLMs for better search retrieval
  Learn how DoorDash uses large language models to improve search accuracy by better understanding complex user queries and matching them to the most relevant items.

---

Data Engineering

• How Notion Brought Order to Its Data Chaos

• 500X Scalability of Experiment Metric Computing with Unified Dynamic Framework
  Learn how Pinterest achieved 500x scalability in experiment metric computation by implementing a Unified Dynamic Framework that eliminates upstream dependencies, accelerates metric delivery and simplifies pipeline development.

• What the Heck is OpenMetadata?
  Learn how OpenMetadata, an open-source platform inspired by Uber's metadata infrastructure, centralizes and simplifies metadata management to enhance data discovery, lineage tracking and governance.

---

Data Analysis and Visualisation

• ChatGPT’s Rising Traffic vs. Other Top Websites
  The rise and rise of ChatGPT traffic

• When the Metric Becomes the Monster
  Learn how Goodhart’s Law quietly wrecks your metrics. When the target becomes the game, teams start optimizing for the number, not the outcome. Here’s how to spot it and what to do instead.

---

"I’m a coder … a vibe coder"

---

As I mentioned last week, if you’re interested in doing a guest post, just reply to this email or message me on Substack and we’ll tee something up.

---

Keep reading: