Products
Stack Overflow for Teams
Capture, share, & collaborate on knowledge internally.
Advertising
Promote your product or service to developers and technologists.
Talent
Engage the world’s technology talent with your employer brand.

Issue 282: No worries about compute

Before we get into the goodies, please take a look at our Developer Survey for this year (link in the footer). It's an annual tradition that builds on the results of previous years to track the tech, habits, and opinions of software technologists. There's more in there about AI this year, but frankly, there's been more AI this year too, and we want to know if it's becoming a part of your workflow or if it's just another hype cycle. XX

Speaking of AI, we've got two conversations about AI and an article about the data behind it. First, we chatted with the DeepMind team about Gemini 2.5, that plucky little GenAI tool that's barnstorming all the leaderboards. On the other hand, we've got a chat with mad scientists chaining together open-source models, compute be damned. If you're in the model training biz, we'd like to talk to you about diversifying your datasets. XX

Now for the newsletter equivalent of speed dating. Could Leia have saved all those Bothans from doom? Is your List of Despair really about the list? Could you maybe go a little easier on the reauth? And rewiring your brain with jams, bops, and power ballads.

From the blog

“We’re not worried about compute anymore”: The future of AI models

Ryan Donovan and Ben Popper sit down with Jamie de Guerre, SVP of Product at Together AI, to discuss the evolving landscape of AI and open-source models. They explore the significance of infrastructure in AI, the differences between open-source and closed-source models, and the ethical considerations surrounding AI technology. Jamie emphasized the importance of leveraging internal data for model training and the need for transparency in AI practices.

Why you need diverse third-party data to deliver trusted AI solutions

Diverse, high-quality data is a prerequisite for reliable, effective, and ethical AI solutions.

Better vibes and vibe coding with Gemini 2.5

Ryan and Ben welcome Tulsee Doshi and Logan Kilpatrick from Google's DeepMind to discuss the advanced capabilities of the new Gemini 2.5.

Caylent Accelerate™ Modernizes Databases 3x Faster with AI

Exclusive first look: Caylent Accelerate™, an AI-driven solution automates up to 70% of the migration process to help enterprises modernize faster—without the risk; breaks vendor lock-in and eliminates licensing costs.

Interesting questions

What happened in the graphics card benchmark scandal in the 1980s?

A corollary to Goodhart's Law: Any metric used as a benchmark becomes a target.

How to make task list less depressing?

"The problem was never the list."

How can historians go to the past without changing it?

Finally, a question relevant to my dissertation.

Doesn't Leia know about the Death Star's weakness already?

Pretty sure she's a diplomat, not a structural engineer.

Links from around the web

Dancing brainwaves: How sound reshapes your brain networks in real time

Oh good, now all the lifehackers are going to optimize their mixtapes.

Frequent reauth doesn't make you more secure

Unless you have a new phone, your apps may not often need to ask "Who dis?"

Trusting your own judgement on ‘AI’ is a huge risk

"It works on my machine" but for your easily fooled brain.

How we decreased GitLab repo backup times from 48 hours to 41 minutes

The great thing about open source is that you can fix your own issues.

There's still time to add your voice to this year's Developer Survey! Take the survey now.