With Project Solara, we are building a new platform purpose-built for agent-first devices.
Excited to work with Cristiano R. Amon and Qualcomm on this! We are in the middle of a platform shift: moving from devices built for apps and operating systems to devices built around agents. It’s changing the nature of computing, from silicon to cloud. In my conversation with Satya Nadella, we talked about what this means for the next era of computing, especially as AI changes how we interact with technology. The opportunity is to build agentic experiences that feel natural, personal, and useful across many form factors. Watch our discussion from #MSBuild.
Read the OnQ blog post for more. [https://lnkd.in/gW6EwiCf](https://lnkd.in/gW6EwiCf?trk=public_post_reshare-text)
[Vishal Show](https://in.linkedin.com/company/vishal-show?trk=public_post_comment_actor-name)3h
The platform plays, but adoption hinges on organizational readiness, not device capability. Agent-centric devices require decision architecture, governance models, and trust frameworks most organizations don't have. The companies extracting value won't be those with better agents. They'll be those that can restructure how decisions actually get made and who controls agent autonomy boundaries.
Ron H.1h Agent-first devices are a bigger shift than they first look. The screen used to be where humans stayed in control. If agents start working inside the device, that control point moves. Then the hard question is not what the device can do. It is what the agent is allowed to see, decide, and trigger. That is where the real operating risk begins.
ADITYA DORGE58m The biggest bottleneck with Enterprise AI today is skyrocketing API token costs and context window bloat. At DestinyAI Multiverse, we have engineered a two-pronged solution to solve this exact problem: Token Saver Engine: It slashes Input & Output tokens by 70% to 90% for complex coding, reading, and writing tasks. What used to cost $100 in API bills drops down to just $10–$20, saving 90% of your context window! Desktop Automation Engine (DAE): A high-accuracy (99%) orchestration layer consisting of 46 subengines and 1,200 modules working like 'one soul' to automate hyper-complex desktop workflows. The Math: Combining 90% cheaper context with 99% accurate automation means a team of just 2 employees can manage workflows that previously required 10 people. We are building the future of autonomous enterprise workflows straight from India. Let's connect! 🔗 Follow our journey or DM me for an early-access look!
A fascinating shift, the technology challenge is no longer just making agents more capable, but making them dependable in real-world environments where context, uncertainty, and consequences matter. as agentic systems move beyond productivity tools and into operational domains, success will increasingly depend on trust, validation, and the ability to demonstrate reliable behaviour under conditions that differ from those seen during development the next platform era may be defined as much by assurance and adoption as by AI capability itself..
Diinesh Bhati10h Project Solara feels like a glimpse into what comes after the app era. For decades, we've interacted with software through screens, apps, menus, and clicks. The vision here is fundamentally different: AI agents that understand context, move across devices, and help complete tasks rather than simply launch applications. What stands out is that Microsoft isn't just building better AI models—it's rethinking the entire computing stack, from devices and operating systems to agents and enterprise workflows. That's a much bigger bet than adding AI features to existing products. If the PC era was defined by apps and the mobile era by touch, the next era may be defined by agents that work alongside us across every form factor. Exciting times ahead for builders, developers, and anyone imagining what human-computer interaction looks like beyond the smartphone. 🚀
Arijit Ghosh15h We spent 30 years building devices for apps. Now we are building devices for agents. The shift sounds technical. It isn't. An app waits to be opened. An agent acts before you ask. That's not a new product. That's a new relationship between humans and machines. 🌏
The shift from app-first to agent-first is the right framing — but it moves the hard problem from the device to the architecture underneath it. When agents act before you ask, the questions that matter are: who authorized that action, what scope did it have, and what happens when it's wrong? The device layer is Qualcomm's problem. The identity, permissioning, and audit layer is every enterprise architect's problem. Project Solara makes that second problem more urgent, not less. Watching how the enterprise security stack evolves around agent-first devices closely.
Project Solara is redefining computing with AI agents that move seamlessly across devices. Excited to see Qualcomm and Microsoft driving this agent‑first future.
Qualcomm14h Looking forward to it. 🤝
Every platform shift creates new opportunities and, sadly, new exclusions. The real test of agent-first computing won't be how intelligent the agents are, but whether everyone can benefit from them regardless of ability/disability, language, age, or technical expertise. Thank you, Satya Nadella and Cristiano R. Amon, for sharing this exciting direction 💓