"If you've got your systems of record, don't screw with them. Keep them and put your agentic around that. If you rewrite a system of record, good luck with an audit. It is like a root canal with a chainsaw," said @dgiambruno at #AIF2026. constellationr.com/insights/… @ldignan @mikeni #AI

ROI matters, but it can't measure the full value of transformation. #AI investments often build capabilities that pay off across multiple technology cycles. A broader Business Value framework helps leaders track adaptability, growth, and sustainable value without abandoning financial accountability. Get the scoop from Esteban Kolsky: bit.ly/4yAnyiw

MyPOV - I asked one of the CRM legends, @DavidSchmaier about his original charter @Salesforce - to build the next generation CRM. David sees a future past agents with autonomous enterprises showing up soon, mentioned hospital etc. In the meantime Salesforce wants to be the trusted platform for the transformatkom. #DF26

Last month I wrote about how we can build a positive and safe future for everyone: meta.com/thefutureisforevery… Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens. The reality is: - People won't want to use agents that are misaligned with them and that don't do what they ask, so labs have a strong natural incentive to make their models more aligned. There is a lot of debate about slowing progress on capabilities until alignment catches up. My view is that trust and alignment are quickly becoming the most important capabilities that will differentiate agents and models. Any lab that doesn't focus on alignment will fall behind. - Labs face significant liability if their models cause harm, so they have a strong incentive to prevent this as well. Meta delayed shipping Muse for several months to focus on safety and security. We didn't call for everyone else to do this before we would. We just did it as part of our day-to-day work because it was clearly the right thing for people and for us. I'm proud of the security foundations we've built. - Engaging independent evaluators and advisors is industry best practice. MSL already does this today in several areas because it helps produce better work. Other labs can just do this too. In general, it would be helpful for there to be a larger and more diverse ecosystem of evaluators. - Committing the significant majority of compute towards serving people rather than racing towards recursive self-improvement is one of the best ways to ensure we develop this technology safely. Meta has made this commitment and other labs can do this as well. I believe the key to building a positive future for everyone is maintaining the right balance of power. This is within our power to do.

.@Salesforce still has pricing work to do bit.ly/4gTglnR Salesforce's #Dreamforce26 conference kicks off this week in San Francisco and rest assured there will be plenty of innovations ranging from Agentforce, Slack and Headless 360 and a broad platform narrative.

The countdown to AI Forum is on. On 9/24, leaders from across business and technology will come together in NYC to tackle the questions that matter most as AI moves from experimentation to action. AI agents. Governance. Accountability. Workforce transformation. Business value. What should AI do, and how do we make sure it delivers? Join us at #AIF2026: bit.ly/4xB10P9

Documents shouldn’t be the final destination for enterprise knowledge. In the age of #AI, they can become active sources of intelligence. Rethink the document lifecycle, rethink knowledge, and put documents back to work to power better decisions, workflows, and outcomes. Read more from @lizkmiller: bit.ly/3Ts1Qyp #EnterpriseAI

MyPOV - @Qualcomm is getting into @AWSCloud - Congrats. Top 3 Takeaways (1) Qualcomm is an attractive partner for the AI era datacenter buildout. Even at a cloud vendor with its own chip design. (2) The expertise in frugal chips for smartphones pays off for frugal racks in the AI era. (3) Qualcomm cannot only offset reduced @Apple revenue, it can grow multiples beyond. And we have not even talked about @Modular yet...

Well played

Why you should arbitrage your AI models like your vendors do - by @ldignan dlvr.it/TVMzLG (via @jonerp)

🔥Ready for what's next in AI? Join the brightest minds at the AI Forum 2026 on September 24 in New York City, NY. Bold talks. Real strategy. No fluff. AI for real-world impact. 👉 bit.ly/4xB10P9 #AIF2026 #AILeadership #GenAI

Agentic AI washing is real. As vendors race for #AI budgets, everything is being labeled “AI.” But Oracle stands out. @holgermu’s latest report examines @Oracle’s AI Database portfolio and finds differentiated, production-ready agentic AI capabilities—with no washing necessary. Read more: bit.ly/4cXIPKL

Quantum computing: Where we stand in 2026 - by @ldignan dlvr.it/TVFRWV (via @jonerp)

On Friday @ 11 AM PT, @DisrupTVShow interviews @briansolis, Head of Global Innovation, @ServiceNow & Co-author of Infinite, Dave Wright, Chief Innovation Officer @ServiceNow and @JohnNosta, author of The Borrowed Mind. Watch here: bit.ly/45wfWRS @ValaAfshar @rwang0 #DisrupTV

Nice to see @AIatMeta back in the open weight model camp with Muse Glimmer. constellationr.com/insights/…

Today we're also opening the weights for Muse Glimmer, a great 30B parameter dense model that can run locally. Soon we'll also release the weights for Muse Spark 1.2, our latest foundation model. Meta is a strong supporter of open source and I'm proud of these releases. Congrats to @alexandr_wang and the MSL team for all your great work on these models.

I have seen a lot of AI reviews this year. Most of them measured the wrong thing. Lines of code. Pull requests. Tokens consumed. Number of forward deployed engineers on site. All input. None of it outcome. Raj Sundaresan talked to Constellation Research about what happens after the demo. Why architecture has to come first. Why optionality matters more than any single model. Why a centralized AI office fails the moment you try to scale it. He puts it plainly. You do not want your knowledge sitting inside one model’s walled garden. None of this is new to us. ALTi Labs. Our work on small language models. Domain Forge. Raj had us building for this about twelve months before it became the conversation. Worth twenty minutes before your next AI steering committee. #tokenmaxxing #FDEmaxxing @raj_sundaresan @Altimetrik @rwang0 @ldignan @ValaAfshar @DisrupTVShow @pfersht @jimitarora @prkelker @boccuzzillc Full blog here altimetrik.com/blog/from-tok…

Everyone is talking about Prompt Engineering, Vibe Coding, Loop Engineering, and now Graph Engineering. The interesting part isn’t that each new paradigm replaces the last. It doesn’t. It wraps it. A loop is simply the smallest possible graph—a single node wired back to itself. Graph engineering doesn’t eliminate loop engineering; it composes thousands of verified loops into coordinated systems. The abstraction layer keeps moving higher. * Prompt Engineering → One request. * Vibe Coding → One conversation. * Loop Engineering → One autonomous system pursuing a goal. * Graph Engineering → Many autonomous systems coordinating toward a larger objective. This progression became remarkably clear over the June–July 2026 discussions across the community—Philipp Steinberger and OpenClaw, Boris Cherny’s Claude Code work, Addy Osmani, and many others. Different vocabulary, same architectural direction. So what comes next? I believe we’re already seeing the beginning of the next abstraction: Meta Engineering (Dynamic Organization Engineering -DOE). Now I have Homer Simspon voice in my head for the day. The graph itself becomes the product. Not just the agents. Not just the workflows. The organization becomes a living system capable of rewriting itself.

Agentic AI is moving beyond copilots into execution. See the vendors on the Cross-Platform Agentic AI ShortList by @rwang0 zurl.co/EAc6z @Accenture @Aera_Technology @Boomi @Capgemini @Cognizant @GoogleCloud @Infosys @Microsoft @moveworks @Salesforce @TCS @Wipro