🧠 Google Just Halved Its Prices

Gemini 3.7 Flash landed this week, just three weeks after Gemini 3.6 Flash. That's an aggressive release cadence even by AI standards, and Google says it's a direct result of developer feedback rather than a scheduled bump.

The model is built for coding, agents, and knowledge work, with real gains in software engineering and web development tasks. The bigger story for your wallet: it launches at half the per-token cost of 3.6 Flash. Read more here.

If you're building anything on Gemini, this is worth testing on playground before the pricing normalizes. It's live now in Google AI Studio.

You can now open a Google Doc, Sheet, or Slide directly inside ChatGPT and work on it side by side with your chat, no more alt-tabbing between windows to check what you're referencing.

This sounds small until you've spent an afternoon bouncing between a ChatGPT tab and a spreadsheet trying to keep both in sync. Now the file just sits next to the conversation.

Anyone who's run a long coding session in Claude Code knows the drill: you hit your usage limit mid-task, and the session just stops until you come back and manually type "continue."

There's now an auto-continue checkbox in Claude Code desktop. Flip it on before you hit your limit, and it'll pick the task back up on its own once your limit resets, no babysitting required. Small feature, but if you've ever lost an overnight run to a limit reset, you'll appreciate this one.

Most AI tools wait for you to ask something before they do anything. Viktor works the opposite way: it lives inside Slack or Microsoft Teams, connects to over 3,200 tools across your stack, and starts proactively suggesting (and doing) work after reading how your team actually operates, things like building reports, spotting problems before you notice them, or shipping a full workflow automation.

It's built for small teams who want something closer to an extra hire than another chatbot tab to check. Worth trying if your team is already living in Slack or Teams and drowning in busywork nobody has time to automate.

Caveman is a plugin for coding tools like Claude Code that compresses the AI's responses into a much shorter, more clipped style to cut down on token usage, and by extension, your bill. The marketed savings figure is a steep 65%, though independent testing on real agentic tasks has found the actual savings run closer to single digits, so treat the headline number as a ceiling, not a guarantee.

Worth a look if you're a developer watching your AI coding costs climb and want to see the difference for yourself.

If you've ever wanted to learn something like storytelling or editing from a specific YouTube creator, Claude can now watch the video and turn what it learns into a reusable skill.

Here's the skill. And here’s a quick demo.

Today is Friday, which means today we go live to build AI workflows that help you improve your work. Join us live, build in real time, and get your questions answered on the spot.

👉Register now!

Can’t make it live? Don’t worry, we’ll send out the recording after the session to everyone who RSVP’ed.

Event details:

📅When: Friday at 10AM PT/1PM ET
📍Where: Virtual, Google Meet