As we reflected on the biggest stories of 2025, one throughline emerged. They all illustrate how enterprises grappled with a central problem: how to get the generative AI tools and agents that promise a new era of business productivity to actually work. www.runtime.news/five-importa...

From last night's Runtime newsletter: A look at the hardware-related announcements from AWS CEO Matt Garman's keynote.

New on Runtime: an interview with Microsoft's Jay Parikh on how building software has changed so much in just a few years. While the broader merits of the generative AI revolution are still up for debate, there's no question that the software-development process will never be the same.

New on @runtime.news: We get a little meta. For $10 a month, Runtime supporters will ensure that our mission can continue. Rather than hide our content behind a paywall or clutter our pages with ads, we believe we can get enough support to keep our independent journalism available to anyone.

New on @runtime.news: A series of incidents in August and early September only highlighted what startups and application developers had been talking about for months: AI reliability falls short of what most businesses expect from their cloud providers, even after last week's massive AWS outage.

New on @runtime.news: The latest in our How We Built It series, this time featuring Mastercard's George Maddaloni. We talked about how Mastercard onboarded generative AI tools across its employees and its current approach to AI agents, where MCP is spurring a lot of internal agent building.

Did anybody else catch Jensen on a hot mic as the press conference started raving about the 1912 cognac he was presumably served in the U.K. this week?

New on @runtime.news: Generative AI apps are forcing app developers to rethink the way users interact with their apps. There have always been apps and workflows that can be controlled with voice commands, but genAI apps open up new opportunities to move beyond the taskbar or the hamburger menu.

From last night's @runtime.news newsletter: GitHub's independence within the Microsoft ecosystem was treasured by company executives in the years following in 2018 acquisition. But that era is over, and GitHub is now just another Microsoft product.

OpenAI is living up to its name again. It hadn't shipped any major LLMs that could claim a degree of openness since 2019's GPT-2. But that just changed with the release of two new open-weight models the company says are as powerful as some of its leading closed models.

It's been a choppy few months for Windsurf. First, news leaked that OpenAI intended to acquire the company for $3 billion. Then Google came along with a billion-dollar acquihire, agreeing to pay $2.4 billion for a non-exclusive license to Windsurf's technology and hire about 40 of its employees.

next time you're in a national park, take some photos of the signage and upload it to help these archivists www.404media.co/save-our-sig...

In this one-on-one interview, Amazon CSO Steve Schmidt tells us about: *Increasing nation-state attacks on tech companies *How generative AI is — and isn't — changing cybersecurity, on both the offensive and defensive sides *How Amazon works to prevent unforced errors in code

New on @runtime.news: Amazon chief security officer Stephen Schmidt talks about the rise in cybersecurity threats from nation-state attackers, why generative AI is better at defense than offense, and the security pros and cons of the rush to develop software with AI tools.

New on @runtime.news: Snowflske CEO Sridhar Ramaswamy increasingly sees data ingestion, analysis, and production as part of a “workflow,” meaning the steps should be easy to follow and more accessible to a wider group of people. Live (kinda) from Snowflake Summit ‘25:

New on @runtime.news : Last year at the Databricks Data & AI Summit CEO Ali Ghodsi pledged to develop a "USB-C format" for data. The industry has made progress toward that goal, but the last mile is tricky, and new tools could let end users worry about other problems.

Don't miss our regular roundup of the week's most important new enterprise tech product releases and updates, including: -Coding agents from OpenAI and Google -AWS's Transform service -Updates to Celonis's Process Intelligence API -Boomi AgentStudio

New on @runtime.news: An interview with Karri Saarinen, co-founder and CEO of Linear. We've been hearing a lot of buzz about Linear over the last couple of years, in part because "we have this fairly simple idea that engineering is really the front line of all this information," he said.

New on @runtime.news: A conversation with ServiceNow CEO Bill McDermott on the sidelines of Knowledge 2025. Highlights include: the impact that AI agents could have on the enterprise, and the ongoing shift where IT departments are taking back control of their application footprints.

New on @runtime.news: An interview with Freshworks CEO Dennis Woodside, who is trying to move the company beyond its SMB roots to take on big IT management and customer-service companies like ServiceNow. We also covered the evolution of agentic AI pricing and Freshworks' potential for expansion.

Synadia and the CNCF have reached a formal agreement to transfer two NATS-related trademarks to the CNCF, after Synadia CEO Derek Collison told Runtime earlier this week it planned to do so. The CNCF's blog post: www.cncf.io/announcement... Earlier on Runtime: www.runtime.news/how-synadias...

For the #rsac crowd, check out the latest edition of the Runtime Roundtable, in which our panelists answer the question: How should companies make sure they are shipping secure generative AI apps while still moving quickly?

New on @runtime.news: A deep dive into the dispute between Synadia and the CNCF over the NATS project. After a contentious weekend, Synadia appears to be standing down. CEO Derek Collison told Runtime Tuesday that it intends to hand over the NATS trademark to the CNCF in the near future.

New on @runtime.news: A deep dive into the dispute between Synadia and the CNCF over the NATS project. After a contentious weekend, Synadia appears to be standing down. CEO Derek Collison told Runtime Tuesday that it intends to hand over the NATS trademark to the CNCF in the near future.

From tonight's @runtime.news: It's been a year since Microsoft rededicated itself to security, and on Monday it released an update. It hasn't dealt with a serious incident since, but Microsoft still faces the inherent tension between protecting customer data and selling enterprise security software.

From last night's @runtime.news newsletter: Google Cloud fell behind AWS and Microsoft during the cloud computing land grab a decade ago, but the greenfield generative AI opportunity has allowed it to reset the playing field. Some thoughts from Cloud Next '25.

Neo4j CEO Emil Eifrem says he sees a recurring pattern in how customers use the database. Due to frustration with SaaS spending paired with increased productivity of in-house engineering teams, he says, "We're moving into—in this classic build-versus-buy (dilemma)—a build era for the enterprise."

“On Friday, when I asked Oracle for comment, a spokesperson asked if they could provide a statement that couldn’t be attributed to Oracle in any way. After I declined, the spokesperson said Oracle would have no comment.”

and just like that CoreWeave has started trading, at $39 per share, below the $40 IPO price. www.cnbc.com/2025/03/28/c...

From last night's @runtime.news: Anthropic's Model Context Protocol (MCP) is "an open standard that enables developers to build secure, two-way connections between their data sources and AI-powered tools." Cloudflare now supports remote MCP servers, a big step toward adoption.

New on @runtime.news : Writer has raised $326 million to build a "full stack" approach to generative AI by developing its own LLMs and building application tools around those models. CEO May Habib thinks using synthetic data and a collaborative process will spur enterprise AI adoption.

From last night's @runtime.news newsletter: Google Cloud's $32 billion deal for Wiz — a cloud security platform designed to work across all major cloud providers — will be a test of its commitment to work with its competitors. Wiz customers on AWS and Azure will be watching closely.

Apps built for or around generative AI have different hardware and software requirements than their older cousins. At the #HumanX conference, we talked to four tech leaders who gave their advice about what to do and what not to do when drawing up infrastructure plans for AI.

Rough day for Salesforce: A massive Slack outage followed by a revenue miss and lowered full-year guidance.

New on @runtime.news: Pinecone just announced an overhaul of its serverless managed vector database infrastructure that focuses on balancing speed, performance, and freshness without making users fiddle with a bunch of dials. My interview with founder and CEO Edo Liberty:

I just wrote a quick blog post about why Substack writers should switch to Ghost! micahflee.com/2025/02/not-...

New on @runtime.news: Our latest How We Built It features Monica Caldas, CIO of Liberty Mutual, explaining how the company quickly rolled out an internal GenAI application thanks to a strong data foundation and FinOps discipline.

New on @runtime.news: Canva CTO Brendan Humphreys explains why it took Canva so long to wade through "gimmicky" internal GenAI tools, and how by giving employees space to try them at their own pace, found the ones that worked. Per-seat licensing is ngmi. www.runtime.news/canva-chose-...

Well, the law review submission cycle is starting up again, so it's time to share what I've been working on: a 2,000-word article on the batshit commercial in which AI overrides Matthew McConaughey's hierarchy of needs: defector.com/salesforce-i...

"Individual engineers experimenting directly on live systems... violates every best practice for controlling risk in the industry. Elon Musk says he wants to run the government like a business, but no business operates this way." Super explanation by @lizthegrey.com of the dangers posed by DOGE.

New on @runtime.news: @vercel.com is introducing a new serverless infrastructure today called Fluid Compute, which makes serverless AI apps more efficient through concurrency. As @rauchg.blue told me, "you're treating it more like a server when you need it." More details here:

New Exclusive building on Wired's reporting: Musk operatives have already pushed live to production extensive code changes to the Treasury Department payment system which makes 95% of fed govs payments. talkingpointsmemo.com/edblog/musk-...

Nvidia claims they're called "NVIDIA" but even Wikipedia doesn't believe it.

New on @runtime.news: A look at a crazy 36 hours in tech, and why DeepSeek shows how constraints can produce an engineering breakthrough that may not be "an extinction-level event," but certainly got Big AI's attention. Powerful but cheap models could solve a lot of problems for enterprise adoption.

This was the best “catch-me-up-quick” thing I read this morning on DeepSeek and the AI stock market, from Stratechery:

Good question: www.datacenterdynamics.com/en/opinions/...

Google auto opted in *every* paying Google Workspace user to Gemini. Also adding a nice increase in price, even though no one opted into it. You cannot turn off / opt out. Except... if you open a support ticket. Only then does Google enable the setting to turn it off! I did:

The latest edition of @runtime.news is out: Why all this AI data center construction feels like dark fiber all over again.