Matt Bernier

You don't invest in features; you invest in architectural shifts that create new economies. The shift from a "read-only" web to an "agentic" web is one of them. Chris Dixon wrote about "Why Decentralization Matters." Sonal Chokshi Robert Hackett Scott Kominers Sam Broner Elizabeth Harkavy and the @a16z crypto podcast have explored how AI agents will redefine interaction. The common thread is the need for a trust and identity layer that is native to this new world, not bolted on from the old one. The problem is, today's internet naming systems (DNS, ENS) are like street addresses for static houses. Agents are not houses; they are dynamic actors who need to move, interact, and transact across the entire web. They need a new kind of OS. At @3ns (Web3 Naming System), we have built that secure operating system. It aggregates fragmented protocols and provides the core, verifiable identity that allows for an interoperable Agent Economy to finally emerge. We have traction, revenue, and a clear path to becoming the default infrastructure for this multi-trillion dollar shift. We're not building a better feature. We're building the architecture for what's next. Comment below if you’d want to see a16z support 3ns #a16z #VentureCapital #Web3 #AI #DeepTech #FutureOfTech #3ns

84

78 Comments

We recently released lifecycle reporting in DX, and I’ve loved seeing it help Platform teams build a concrete business case for funding important Platform and DevEx initiatives. While traditional metrics—like build times, pull request merge times, and issue cycle times—offer useful insights, they cover only a narrow slice of developer workflows. They also often fail to distinguish between non-blocking delays (where developers can switch context and stay productive) and true friction points that drain valuable engineering time. Lifecycle reporting helps pinpoint where in the development process teams are experiencing the most friction and delays—and ultimately costing organizations the most. Just like in the screenshot below, I’m seeing companies lose thousands of engineering hours each year to challenges understanding existing code, lengthy CI processes, dev environment issues, and excessive meeting schedules.

123

6 Comments

So which models produce HTML that has the least accessibility errors, and at what cost? Mapped it out here (bottom left is the best). This is helpful for understanding which models are most compliant for those with disabilities and do it efficiently. Qwen and Gemini 2.5 Flash lead the pack. Hat tip to Ben Ogilvie for pointing me to this: aimac.ai

50

2 Comments

📅 Today at revAlation Chicago, we’re unveiling Alation Agent Builder— enabling you to build accurate, configurable, governed, metadata-aware agents. Prototypes are easy. Accuracy and trust are hard. Especially with structured data, where even a small error in a query or filter can break trust. Agent Builder helps teams: 🚀 Reach 90%+ accuracy before deployment with built-in evaluation tools 🛠️ Build agents with no code and prebuilt templates 🤝 Connect to 100+ data sources and apps 🏢 JLL is already using it to power intelligent lease decisions with trusted data (and they're speaking at revAlation today!). https://lnkd.in/g58AeF7u

105

5 Comments

Adding Document Understanding to Claude Code 📑🤖 Coding agents are awesome but they don’t natively understand docs (they kind of do, but it sucks). This limits their ability to build business applications - think about anything from a financial copilot to contract reviewer to rfp synthesizer to synthesizing PRDs. I wrote a blog to capture core patterns for adding document understanding to Claude Code - applicable to coding agents in general. This both helps Claude Code 1) get enterprise context when building apps, and 2) allows it to build more agentic, robust apps in general. 1️⃣ Add an MCP server that preindexes your file knowledge base 2️⃣ Let Claude Code loose on your filesystem with the right CLI tools for understanding files. 3️⃣ Give Claude Code context on core modules for document understanding and agentic workflows - this last step is important if you want to build “better apps in general” and not just “rule-based apps”. Blog: https://lnkd.in/g4EEFftk The blog offers general principles regardless of whether you’re using our stuff, because we fundamentally think that coding agents + docs leads to a ton of useful applications. Of course we are deeply investing in tooling in this space too (some referenced in the blog), a lot of it dependent on LlamaCloud. Check it out: https://lnkd.in/g9Wpqn7w

210

3 Comments

At LoopQA, like most firms, we start with a POC to show what we can deliver. There are countless metrics you could use to measure the impact of augmented coding, but let’s just look at our own POC process as an example. We build a basic smoke/sanity suite with a page object architecture using Playwright. Normally, that would take about a week turnaround, 10–15 hours of total work. With augmented coding, we turn around that same project in just 2–3 hours. The first objection is usually: “Well, of course it works for simple things.” And yeah, this is a straightforward build project. But it works just as effectively for refactoring, documenting, and cleaning as it does for greenfield builds. ** As a note, I know I've fallen off on YouTube. I've been slammed. For those asking about more example videos, I'll be starting those back up next week.

29

5 Comments

ChatGPT-5’s release this week isn’t just a product update It’s the clearest signal yet, AI is crossing the line from tool… to team member We are now in the era of Agentic AI, systems that can plan, decide, and execute without constant human prompts. We often talk about AI like it’s a software upgrade But what if it’s your next workforce? The biggest misconception about AI today is treating it as a one-off automation project, something to make things “faster” or “cheaper” That’s not transformation. That’s optimization The leaders who will win this decade aren’t just deploying AI to cut costs. They are restructuring operations around intelligent, autonomous agents that can own outcomes, work across systems, and scale impact far beyond human capacity This isn’t about replacing people It’s about building AI-powered teams that work alongside your best talent—unlocking new revenue streams, compressing delivery timelines, and creating competitive advantages your rivals can’t match If you’re a CEO or senior leader wondering how to turn AI from a buzzword into a bottom-line driver, now is the moment to act I’m open to conversations with leaders ready to explore what Agentic AI could mean for their business, whether you’re starting from zero or scaling your next big move. #AgenticAI #AIForBusiness #AILeadership #EnterpriseAI #FutureOfWork #AITransformation #AIAdoption #NextGenAI #GenAI

33

2 Comments

Ebooks are dead. At least, that’s what I used to think. (Just kidding… sort of.) Then we spent months talking to producers, compiling data, and putting together something that wasn’t just filler content. This is the first ebook we’ve done at Slabstack. And it’s built for suppliers who’ve outgrown spreadsheet quoting, are sick of pricing leaks, and want to finally bring margin discipline into their sales process. We packed it with stories from teams on the ground, breakdowns of real quoting errors, and frameworks we’ve seen actually work. If you’re trying to move from quoting fast to quoting smart, this is for you. Get your full copy here: https://lnkd.in/eXrDrqrQ

25

3 Comments

We just dropped Episode 1 of The DevGTM Craft & Orlando Nieves absolutely delivered!! Military intelligence → RevOps at Unstructured → Converting mass adoption of their open-source software into timed and contextual enterprise pipeline. This is exactly the kind of DevGTM kungfu Reo.Dev was built for! Orlando Nieves, Revenue Operations Lead at Unstructured talks about how they: -- Turn intent signals into custom landing pages that grab buyer attention -- Use AI agents to predict perfect outreach timing based on pipeline patterns -- Build an enterprise business case based on prospects’ own OS usage data. I really liked this part: "We want to get to that sweet spot of reaching out at the right time... We don't want them to already be committed to that rat's nest." That's it. That's the entire DevGTM game right there. Too early = no problem awareness. Too late = they're married to their homegrown solution. Orlando gets it. And now you can too. 🎙️ Listen to the full episode on Spotify, YouTube, Apple. [Linking in comments below] This one's for the operators who know that great products need great go-to-market. #DevToolGTM #IntentSignals #OpenSourceGTM #BuildInPublic #DevGTMCraft

20

1 Comment

What are MCPs and why should AI Operators pay attention? APIs were how we connected systems. MCPs are how we’ll connect agents. If you’re new to MCPs - think of them like a smarter API. Instead of needing exact instructions, you can tell it what you want (“a 4-course pasta dinner”) … and it figures out the best way to make it happen. That’s what makes them powerful for AI Operators. If you already run your work through a clear playbook, MCPs just give you a smarter way to run a step. It's not a fancy new game-changing you-must-drop-everything-and-switch-to-this tool like social media may make you think Just a faster way for your agents to do the work you define. What's nice is that they are what we call "modular" - meaning once the MCP is built, any other agent can plug into it and run instantly. Think of it like giving your agents a shared “language” for executing your playbooks. Have you dabbled in MCPs yet? What questions do you have about them?

45

12 Comments

How to give engineers requirements in 2025: (Hint: it’s not a ticket that says “Add a search bar”) Start with the why: “Our users are frustrated because they can’t find anything once their data grows past a few dozen items. Support tickets for this jumped 40% last month.” Define the scope: “This needs to work for both mobile and desktop. We’re not optimizing for advanced filters yet, just basic keyword lookup across titles and tags.” Give clarity: “The goal is to reduce time-to-discovery. Success means users can find what they need in under 5 seconds. We’ll measure success through session recordings and search success rate.” Reinforce focus: "Speed and simplicity matter most. We’d rather ship a fast, usable v1 than a perfect UX.” Then step back. Let your engineers own the how. Not just build what was asked. Because real requirements in 2025 aren’t just specs. They’re context. They’re constraints. They’re customer outcomes. That’s what it means to be Product Driven and that's exactly what we do at Algorithm

51

11 Comments

Private AI is not coming. It's here. And it's more than capable of doing 70 percent of enterprise work today. That includes complex reasoning, code generation, content workflows, RFP response, support, summarization, research, BI tasks, and more. These aren't speculative use cases. They are repeatable, testable, and automatable using open models and agentic systems already available. The latest results from the Qwen team prove this plainly. Their new release, Qwen3-235B-A22B-Instruct-2507, shows top-tier performance across some of the most demanding reasoning, logic, and coding benchmarks, including GPQA, AIME25, LiveCodeBench, Arena-Hard, and BFCL. In key cases, it's outperforming Claude Opus 4 and matching or exceeding models like DeepSeek-V3. What makes this different is that it is running in private mode. No server calls, no external logging, no vendor lock-in. It is a blueprint for what enterprises need right now models that are fast, aligned, secure, and extensible. Qwen3 isn't alone. The broader ecosystem of open-source models is evolving faster than most realize. Every release brings better alignment, longer context handling, broader language support, and more accurate reasoning. These capabilities are landing right as enterprises begin to integrate multi-agent orchestration, private inference mesh architectures, and fine-tuned vertical systems into their daily operations. If you’re building the next chapter of your business and you’re still waiting for AGI, you’re missing what’s already here. The 70 percent of tasks that can be done today don’t need general intelligence. They need reliable autonomy. They need architecture. They need scale. And now they have it. The future of knowledge work is not someday. It is already being executed inside private stacks. Agents are being deployed to reduce toil, increase insight, and drive velocity across critical functions. Private AI is not a research project anymore. It is a productivity movement. The only question left is how fast you will move. #PrivateAI #EnterpriseAI #GenAI #OpenSourceModels #Qwen #LLMs #AIInfrastructure #AIProductivity #KnowledgeWork #Agents #RAG #InferenceMesh #AIatWork KamiwazaAI

25

2 Comments