Looking back through today’s list, there’s something that jumped out at me. You should ask a question after reading the headline. Why didn’t OKRs work? Why are predictability and outcomes at odds? How many pilots of any sort make it all the way to production? Why rewrite all your code? Let’s be good question-askers, and seek answers.

[blog] “You didn’t do it right.” Every measurement framework isn’t right for every team. Jason and team tried OKRs, and they just didn’t land. It’s often not a knowledge problem, but just the wrong context for it to work. That’s ok.

[article] 38 things we’ve learned about building AI products. Sheesh, what a fantastic list of lessons learned. You’ll have your own, but digest these too.

[blog] One Agent Development Kit, five languages. If you’re going to build a code-based agent—and I’m not positive that’ll be the dominant way as we see more pre-built, general agents that get fed custom skills instead—it’s useful to have a common framework across programming languages.

[blog] Predictability vs Outcomes. This made me stop and think. It’s not output versus outcomes, but the decision between predictability and outcomes.

[article] How Anthropic scales telemetry and billing data with Pub/Sub. Pretty wild stats, and a testament to using an at-scale cloud messaging engine.

[blog] How to Manage Context in AI Agents. This all feels very unsettled to me. Shir does a good job articulating the options. But I do see people debating the merits of all these approaches!

[blog] Previewing Gemini 4 Argon by Building an AR Game for My 4-Year-Old. Thats a tough customer. How great is it that we can realize our ideas so quickly now?

[article] Why 88% of AI pilots never make it to production. Noted. But also recognize that 20-30% of ALL pilots make it production. It’s hard for technology to go from a perfectly curated clean room to the rough-and-tumble production world.

[blog] 6 Years of Tech Debt, 1 Prompt: Modernizing a Legacy App with Gemini 4. Excellent demo scenario from Mete. A six-year full-stack upgrade in one shot? Remarkable.

[blog] How to effectively debug when you no longer understand the code. A small novel in length, this piece from Cindy is a wonderful look at expertise, assumptions, and how to rethink our directions to AI agents.

[article] It’s time to rewrite all that old code. Sure. But first ask “why?” And “in what language?” And “which tool?” Maybe “targeting which platform?”

[blog] Better HTTP/1 connection reuse. Geeky topic? Yes. But I like knowing my coding language takes care of stuff like this on my behalf.

[blog] Predictable AI spend in Big Query: Introducing cost management for generative AI functions. I like being able to embed LLM calls inside my queries, but it’s probably easy to go overboard. I’m glad we have quotas in place for you.

Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:

Llama 3 is a capable generalist, but that's exactly the problem when you need an agent that reliably outputs structured JSON for a specific API.

"Starting today, Linux desktops are now officially supported by Microsoft Defender," reports Neowin. A public preview of the threat detection app extends its enterprise security platform beyond servers to developer workstations running distributions such as Ubuntu, Debian, Fedora, and RHEL. The offering includes antivirus protection, endpoint detection and response, vulnerability management, threat hunting, and device isolation. From the report: Microsoft is also making it easier to onboard desktops through streamlined methods, offering a unified security operations experience, strengthened security, and improved response times. Organizations that are already familiar with securing Linux servers via Defender should feel right at home securing desktops too, as the experience is exactly the same. Defender knows whether it is securing a device in a data center, the cloud, or that it is owned by a developer and identifies them as such in inventory and security workflows. As stated previously, all Linux server distros are supported via Defender on desktops too. These include RHEL, CentOS, Ubuntu, Debian, SLES, Fedora, and Oracle Linux, GNOME, KDE, and more. That said, you do need Microsoft Defender for Endpoint Plan 2 or a Microsoft 365 E5 and E7 subscription. You can enroll Defender on workstations using the same deployment methods you use for servers, but do check out the prerequisites here.

Read more of this story at Slashdot.

It's been four long years since the last Apple TV 4K was released, and in that time there have been lots of hopes and dreams for features in whatever comes next. A new chip, 4K/120Hz support, audio passthrough, a remote locator, better home control integration - the 2022 Apple TV 4K is a great box, but there's still plenty of room for improvement. With the fourth-gen Apple TV 4K announced on Tuesday, some of those hopes were met, while others were unfortunately left out.

Coloring the announcement is the device's price. Apple eliminated the cheaper 64GB, Wi-Fi-only model, leaving just the 128GB, Wi-Fi and Ethernet model - and bumped up its p …

Read the full story at The Verge.

If you don't get AppleCare Plus for the iPhone Duo, any repairs are going to be pretty expensive. Apple published the estimated repair costs for the iPhone Duo without AppleCare Plus today, as reported by MacRumors, and they can be high: Battery service, for example, will cost $229, inner display damage will cost $999, and "other damage" costs $1,499.

Here's Apple's estimated costs for various repairs if you don't have AppleCare Plus:

Cover layer damage $49

Back glass damage $199

Battery service $229

Rear camera damage $229

Outer display damage $379

Outer display and back glass damage $509

Inner display damage $999

Ot …

Read the full story at The Verge.

Following a report last week that Netflix was planning to lay off five percent of the company, today the streamer's co-CEOs Ted Sarandos and Greg Peters announced in an email that it would actually be laying off six percent of its workforce, or around 1,000 employees, according to Variety, which obtained a copy of the memo. That makes this Netflix's largest round of layoffs so far, following the company cutting 150 jobs in May 2022, and an additional 300 in June 2022 in an attempt to reduce costs following slower revenue growth.

Today's layoffs will mostly affect staff in Netflix's product, technology, human resources, and legal teams, acco …

Read the full story at The Verge.