IV.AI

March '26 in AI was MADNESS even for short people. Here’s a RECAP *** General Stuff *** Governments are accelerating AI policy. The EU AI Act is in force with tiered compliance and the White House released its own National AI Legislative Framework. AI plays an active role in the Iran conflict, supporting intelligence, surveillance, and information operations. Open-source orchestration is getting serious with tools like Ruflo spinning up 60+ specialized agents in coordinated swarms. The “ChatGPT moment” for robotics is here with NVIDIA pushing Physical AI at GTC OpenAI shut Sora down and $1BN Disney deal collapsed Two teenagers built a “vibe-coded” AI calorie-tracking app and sold it for ~$80M A man used ChatGPT to design a personalized mRNA cancer vaccine for his dying dog, Rosie. Model race: GPT-5.4, Gemini 3.1, Grok 4.20, and Mistral Small 4 all dropped in a 23-day window. MCP hits 97 million installs Quiet infrastructure milestone. OpenAI crosses $25B ARR. Anthropic at $19B. Both companies are approaching IPO territory. Google drops Lyria 3 Pro DeepMind's most advanced music model, live for paid Gemini subscribers. Music generation is quietly getting serious. Atlassian cuts 10% of staff, replaces CTO with two AI-focused CTOs 1,600 people gone. China's agent ecosystem Tencent quietly launched a full agent product family and pushed it into WeChat. The West barely noticed. The OpenClaw wave is global. Then today, just today, Anthropic accidentally leaked their entire source code to the public. All links in Comments *** Big Theme *** AI at War Anthropic said no to powering autonomous weapons systems and surveillance for the Pentagon. The DoD called them a “supply-chain risk.” The Trump administration ordered a federal phase-out. OpenAI then swoops in and says yes. Public backlash followed. OpenAI then amended the deal with explicit domestic surveillance protections. Despite all the drama, the military remained dependent on Claude. That’s in the period of just 10 days! *** 2nd Big Theme *** Agents and tools explodes further On the back of the viral OpenClaw in Jan / Feb, March was the month of the big players jumping in with their own additions. Claude: scheduled tasks, voice mode, daily model upgrades. Then Anthropic rolled out Claude Cowork, Dispatch, full computer control. NVIDIA dropped NemoClaw: enterprise agent infrastructure, open-source Perplexity launches Computer, running on a local machine ByteDance’s DeerFlow 2.0 lets agents spawn agents and orchestrate entire workflows.   *** Things my human brain thinks *** I think the overarching theme is ‘taste’ As tools get better and the need for razor sharp strategy expands, taste with the ability to move incredible fast becomes a super power.

I'm excited to share that I've joined IV.AI as a Full Stack Developer 🚀 From my very first conversations with the team, it was clear that this is a place where strong ideas, thoughtful engineering, and real world impact come together. I'm looking forward to building, learning, and contributing alongside such a talented group. A special thank you to Vince Lynch and John Martin for the trust and opportunity to join the technical team. I truly appreciate the open door and the vision you've shared. Excited for what's ahead and to start shipping great things together. #NewRole #FullStackDeveloper #AI #Tech

I’m happy to share that I’m starting a new position as Senior Full Stack Software Engineer at IV.AI!

One month, Two AI triggers. Software stocks explode... Let's call it what it is... apocalypstick You may remember the one from Feb 3rd when Anthropic's agent release hit Salesforce, Monday.com, etc with the market dropping about 5% Yesterday, Citrini Research released a hypothetical scenario where essentially most of the white collar jobs go away in 2028 and it triggers a mortgage crisis. That same day in 2026... IBM drops 13%, Datadog 11%, and CrowdStrike 10%. Generally a SaaS drom of about 4% Let’s unpack that. Could AI reshape jobs? Yes it will. But it doesn't mean all these workers will be going to be jobless and defaulting on their mortgages The scenario misses how AI actually works Especially in white-collar work. Roles will change, but blanket job loss predictions ignore how hard it is to make great AI systems What I think is actually happening is these 'shocks' are rightsizing valuations and we're seeing a correction. Not that AI isn't awesome, but it's a lot of work. Human work. AI is a tool Neither apocaclypse nor miracle. Maybe we frame it as apocalypstick? We'll come out the other side with a new look and those who are targeted in their AI beauty routine might be able to buy a house at a reasonable rate? Correction or crash?

AI is growing, so is our team. Welcome Alie C., our new Chief Revenue Officer, to the IV family! With 10+ years at the intersection of sales, data, and tech, Alie is an absolute powerhouse. Her razor-sharp go-to-market mindset and proven ability to solve customer challenges are going to drive IV.AI's next chapter of growth! Right now, the problem with the data industry is compromise. You either get a beautiful dashboard that shows basic, surface-level metrics you already know... or a massive, clunky data dump that requires a team of PHDs. At IV, we focus on the AI solution. Our platform is a funnel, taking your KPIs as the P0, layering in the data you didn't even know you were missing, running it through our secure & customized AI engine, and delivering all of it through an interface that your team will actually enjoy using. We give you enterprise data power with consumer simplicity. Interested in working at IV? Explore our open roles at https://iv.ai/careers

🦾 I’m thrilled to announce that I have joined IV.AI as Chief Revenue Officer. In a noisy market, it is rare to find a team that has been successfully deploying AI solutions for over 10 years. The talent here is incredible and the technology is truly solving complex problems. If you are looking for a personalized AI strategy, or need to access data that seems impossible to find—reach out. I’d love to show you what we are building. Big shout out to Arteen Arabshahi for connecting me with Vince, and to Chris Denson for being the best innovation mentor in the game. #AI #ilovemyjob

The extreme interest in ClawedBot (now MoltBot because Anthropic apparently hates free high-risk brand building)  https://lnkd.in/gqncB6K8   makes something very clear. automation as a trend explodes fast when it hits the pings are deafening automation broke containment as people didn't watch AI take goals / complete tasks / navigate systems / automate from the comfort of a mac mini. the risk is real for users with massive potential attack surfaces but so is the volume of opportunity for folks who sell things on the internet. ACP / UCP / SCP suddenly matter even more: https://lnkd.in/gucqvUWF https://ucp.dev/ https://lnkd.in/gVqTBKnU Any business that wants to engage with autonomous agents needs to have the infrastructure layer together Autonomous agents need explicit permissions. What can be read. What can be written. What can be executed. What requires confirmation. ACP and UCP can act as control planes for machine action. Without them, agents reverse engineer behavior through UI scraping and open web, brittle and often inaccurate. Then comes context. Autonomous work without context produces repetition and drift. This is part of why Clawdddd ahem, Molty has grown so quickly Agents forget constraints. They ask questions to known answers which infuriates the spoiled optimizer. SCP exists to make context portable, scoped, and revocable. Who is acting. On whose behalf. With what constraints. Across which systems. ClawedBot carried open source task state. SCP can generalize into durable context across tools, agents, and sessions. So the priority is straightforward. First, define machine level permissions for action. Second, expose capabilities instead of relying on UI side effects. Third, decide what context can travel with an agent and under what rules. Fourth, instrument everything so autonomy can be audited and interrupted. ClawedBot is a stress test. The room is buzzing. Autonomous AI is going mainstream now. Ring the alarm.

The future of retail AI isn’t just about the tech. It’s about the standards we set today. We are thrilled to be a sponsor for the Retail AI Council Industry Reception on January 12th in NYC. This event is a major milestone for the industry because it marks the official launch of the Shopper Context Protocol (SCP). We’ll be joining senior leaders and innovators during #NRF2026 to discuss how these new standards will shape the next era of agentic commerce. Vince Lynch (CEO of IV.AI) and some of our leadership team will be there, and we’d love to have you join us. Request an invite here: https://bit.ly/498Uc12 PS: If you’re curious about how this type of approach has helped AI and user alignment in the past, check out the Open Standard we created with 25K members of the AI community back in 2023: https://lnkd.in/dSajgVc4

2026. The year everyone gets it: In order for an LLM to perform well it needs CONTEXT So, how do you give them context at scale without turning every AI into a thirsty walled garden? The Shopper Context Protocol. An open standard that lets AI assistants securely carry your customer's loyalty status, preferences, and purchase history across conversations and platforms so retailers can maintain personalized relationships even when agents, ChatGPT or Claude etc. Join us as we launch at NRF on Monday at our happy hour:  https://luma.com/h0v1rt31 How it works SCP defines standardized schemas and exchange patterns that sit between AI agents and commerce systems, enabling loyalty signals, preferences, scoped history, and declared intent to persist across sessions and channels Emerging agentic commerce and checkout protocols handle discovery and transactions, SCP provides the missing personalization and continuity layer Shoppers explicitly control what context is shared and with whom. Brands can recognize customers through scoped access without surrendering data ownership or security Provides MCP-compatible interfaces, allowing shopper context to be consumed cleanly by modern AI systems The Stack: Open specification published on GitHub TypeScript-based reference server implementations Lightweight, vendor-agnostic architecture with no proprietary lock-in Secure, scoped authentication and authorization patterns This time of a system will be integral to consumer engagement w/ LLMs in the mix but will take input from everyone. More details about the approach and how you can join the working group here: https://lnkd.in/gVqTBKnU Join us at the Retail AI Council Reception (Monday): This marks the official launch of the Council's Shopper Context Protocol (SCP) Working Group. Doug Weich, Matthew Holland, and Vince Lynch will share how this open standard will make shopper context portable as commerce becomes increasingly led by agents.