Web3 Firehose

Three years ago, I planned for 30 or 40 people at the first Boston Generative AI Meetup. It sold out at capacity at 320. Last night, we crossed 6,400 members over 22 events. Thank you to everyone who keeps showing up. Another packed room. Another honest conversation. Another reminder that Boston has one of the most engaged AI communities anywhere. When I started the Boston Generative AI Meetup, AI was not yet the center of every boardroom conversation. The goal was simple: Bring people together. Create a place for serious conversation. Help make Massachusetts the leading AI economy by 2035. That was the mission I stated at the first meetup. It is still the mission now. The more of us building this ecosystem, inclusively, the better. Last night’s panel was a good example of why this keeps working. Jenn Azar brought the perspective of an operating CEO applying AI inside a 400-person company serving critical industries, including nuclear. She was excellent on what organizational change actually takes and how companies have to evolve into AI, not just adopt it. Doug Levin brought the perspective of a software maker’s software maker, with deep experience building technology companies and thinking seriously about trust, security, and risk. Mike Grandinetti brought the global educator’s view on how AI reshapes organizations, leadership, entrepreneurship, and adoption. Paul Baier grounded the conversation in what companies are actually doing to adopt, govern, educate, and scale AI. Dave Peak showed how real AI agents are already working for small and medium-sized businesses, not as theory, but in live market application. My biggest takeaway: The organizations getting AI right the ones willing to restructure how work actually gets done. Most orgs are still far earlier than the headlines suggest. And that gap - between the noise and the work - is the opportunity. Grateful to Microsoft for the enduring partnership, to Squark AI for supporting the community, and to everyone who shows up, asks questions, challenges ideas, and makes the conversation better. We will post the next Boston Generative AI Meetup in the future. See you there & at Boston AI Week.

Some states are adopting AI at nearly double the rate of others. This map shows America's emerging AI divide.

Explore the U.S. economy in one chart, showing what industries contributed to America's $31.4 trillion GDP in 2025.

CNBC's Kate Rooney reports on news regarding the Elon Musk v. Sam Altman trial.

We built cosmos™because most companies run on systems built for a different era.

One of the questions my Battery Ventures partner Bill Binch and I get most often is: How should I evolve my tech stack with this new wave of AI tools? To answer that question, Bill and I created this "AI-first GTM Tech Stack" graphic. Future versions are already in discussion -- there's definitely going to be more than one way to meet this need. | 62 comments on LinkedIn

ABM is back. And it's all about signals. There are 3 layers of signals every GTM team in 2026 should track: 1️⃣ 1st-Party Signals (Data you own) → CRM- Data already in your CRM. [HubSpot, Salesforce, Attio] → Product Usage- Product analytics triggers. [Mixpanel, Amplitude, Intercom] → Meeting Forms- Demo bookings or form drop-offs. [Chili Piper, Default, Calendly] → Gated Content- Opt-in content/lead magnets. [Webflow, Distribute.so, Gamma] → Marketing Sequences- Engagement with email flows. [beehiiv, Customer.io, Kit] → Webinar Attendance- Attendees of online events. [Luma, LinkedIn events, Goldcast] → Website Visitors- Deanonymized web traffic. [Instantly.ai, Warmly, RB2B] → Outreach Replies- Sales engagement tool prospects. [Instantly.ai, HeyReach.io, OutboundSync] 2️⃣ 2nd-Party Signals (Exclusive data from external tools) → Partner Signals- Shared data from integration partners. [Crossbeam, PartnerStack, WorkSpan] → Warm Intros- Common connections with prospects. [Commsor, The Swarm 🔆, LinkedIn] → Review Sites- Intent from software review sites. [G2, Capterra, TrustRadius] → Champion Tracking- Previous customer job changes. [LoneScale, Champify, Clay] → Ad Insights- Engagement from ad campaigns. [ZenABM, Fibbler, Vector 👻] → LinkedIn Engagement- Interactions on LI posts. [Clay, Jungler, Teamfluence™] 3️⃣ 3rd-Party Signals (Public data from external sources) → Technographic Data- Changes in tech stack. [BuiltWith, HG Insights, TheirStack] → Job Openings- New postings on job platforms. [Clay, PredictLeads, People Data Labs] → Firmographic Data- Company demographic changes. [Apollo, AI Ark, BetterContact] → People Data- Staffing changes. [Findymail, Clay, Apollo] → Job Changes - New hires. [Sales Nav, Clay, Apollo] → News- Company announcements. [Clay, Google News, Perplexity] → Advertisement Activity- Media buying behavior. [Adyntel, Adbeat, Apify] → Web Data- Data scraped from the web. [Clay, Apify, ZenRows] → Social Signals- Engagement on socials. [Clay, PhantomBuster, Jungler] → Search Analytics- SEO changes. [SemRush, Similarweb, Ahrefs] → Funding- Recent fundraising events. [Crunchbase, Owler, Pitchbook] ___ Every marketer wants to know: - Who knows they have a problem? - Who already knows our company? - Who's actively considering buying? Signal stacking is the closest thing we have to answer these questions. Follow Fivos Aresti for more GTM playbooks. | 29 comments on LinkedIn

B2B marketing services are at an inflection point. What was once a fragmented ecosystem of creative agencies, demand-gen shops, digital firms, marketing automation consultants, and analytics boutiques is colliding with a new reality: buyers now expect full-funnel precision, speed, measurable revenue

Every board meeting I attend now starts with the same question:

Early on, you can often just sell a service as your "MVP" - rather than spending a bunch of time building a product that people may or may not want.

From industry-standard to open-source freedom

I’ve been disappointed with HBR articles on LLMs of late and been openly critical. This one’s a notable corrective. Credit where it’s due, The Gen AI Playbook for Organizations is a far more credible, grounded lens. It’s aware of epistemic limits in a way most recent HBR AI pieces simply haven’t been. The authors introduce two genuinely useful axes - cost of errors and type of knowledge (explicit vs tacit) - to decide where Gen AI can be used well. That simple distinction reframes the whole conversation from “how smart is the model?” to “what kind of knowing does this task demand, and what’s the price of being wrong?” It’s still classic HBR - managerial tone, quadrant chart, actionable checklists - but it earns the framework. Think of it as McKinsey meets tacit knowledge, competently. What it gets right: - Reintroduces human judgment: empathy, ethical reasoning, and context as irreducible. - Offers a practical map of where Gen AI belongs today and where it doesn’t. - Frames the paradox of access: when everyone can use the same tools, advantage shifts to how you use them. - Grounds strategy in difference, not speed. What leaders can take from it: If you’re wrestling with Gen AI adoption, three questions here are worth carrying into the boardroom: 1️) Where is the cost of being wrong low enough to let Gen AI run? Start there. Don’t chase perfection; chase learning. 2️) Which tasks rely on tacit judgment, context, or ethics? Keep humans squarely in the loop, that’s where trust and reputation live. 3️) How can we use the time Gen AI saves to improve coordination, not just efficiency? Free hours mean nothing if feedback loops stay broken. Those three prompts alone could help many reshape how leadership teams approach pilots, governance, and ROI. The piece reveals where corporate AI thinking still sits, caught between epistemic humility and managerial control. Where it falls short: - They echo Polanyi without ever naming him, invoking tacit knowledge while designing as if it could be made explicit. - They say strategy beats speed, then prescribe speed as strategy. - Their “cost of errors” model assumes organisations know that cost, when it’s often emergent, reputational, or discovered too late. - They keep humans in the loop but forget humans are the loop, the source of coherence itself. That’s where it overlaps - and diverges - from my own positioning. Where they speak of knowledge types, I speak of rails: the organisational feedback systems, semantics, and judgment structures that turn intelligence into ROI. Where they stop at “human oversight,” I push toward management discipline over model fascination. Still, it’s a step in the right direction, for once, an HBR article on AI that recognises strategy is about where humans remain essential, not how fast machines improve. | 15 comments on LinkedIn

The AI consulting space is full of NOISE - but there’s a small group of firms actually doing BRILLIANT work. Here’s who I’d keep on your radar 👇 Quantum Rise – I'm a big fan of their “Consulting 2.0” mantra: profit, not PowerPoints - frankly, the entire industry needs to move in this direction, but that's another rant for another time. Alex Kelleher (former Deloitte Global CMO and serial entrepreneur) has built one of the earlier - and larger - players in this space with cross-industry impact, especially for companies that operate in the "physical world". Tenex – Alex Lieberman made waves in the consulting community earlier this year when he announced his intent for Tenex to be the “McKinsey for AI.” They're a powerful engineering shop for sure, and I’m particularly impressed by how an “outsider” (aka, not founded by a former consultant) approached this. Goes to show there’s real diversity of thought in this space, and a ton for the rest of us to learn too. Casper Studios – I’ve admired these guys for a while now, largely for their versatility, working with clients from hedge funds to nonprofits to creators. One of the best design-oriented build firms I’ve seen. Jay Singh's approach to design-driven AI is something I’ve learned a lot from. Morningside AI – Honestly, I didn’t think a firm born from a YouTube influencer would ever stand ground against “established” firms - but they’re quickly ceding it. A testament to the team and the credibility they’ve built. Based in New Zealand, recently brought on Nick Roco (another one to watch) to head up their consulting arm. Pathfindr – Also from “the land down unda” (can I even say that?), I’ve been amazed by their practical, anti-consulting approach to AI. They truly make it SIMPLE - exactly what leaders want. Dawid Naude (ex-ACN) has been building something special, and it shows: they were recently acquired by the Affinda Group, a testament to their concept + execution. AI Momentum Partners (AMP) – Word on the street is they have the coolest, funniest, smartest, most handsome founder on the block (Just kidding - I’m way more self-deprecating than that!). AMP is among the fastest-growing strategy-to-execution partners in the AI consulting arena, leveraging an "MBB meets FAANG" approach across service-oriented mid-market and private equity-backed firms. Superposition - Not really an “AI consulting firm” (but rather, a strategy partner for Data & AI companies). Honestly, I just like David Cohen. Sooo... who did I miss? Tag the founders and firms you think are ACTUALLY pushing the AI consulting industry forward - I’ll feature a few in a follow-up post next week. :) #AI #Consulting #AIConsulting #MBB #Big4 #DigitalTransformation | 172 comments on LinkedIn

AI will replace jobs and do so quickly... Anthropic released a study on Monday that showed most businesses (77%) are using its Claude chatbot to automate work. Only 12% of businesses used Claude as a collaborative or learning tool. The report stated that the trend towards automation could “bring disruption in labor markets” and cause workers to lose their jobs And today from Digiday... S4 Capital has dropped the pretense. The company’s top brass are now openly framing AI agents not as enhancement to agency work, but as the replacement. In their view, the traditional agency model isn’t evolving – it’s being automated out. And that automation is already reshaping how ads get made. Clients like Google are now paying S4 for what its AI agents can deliver, not its people. Take a recent Pixel smartphone campaign. AI handled nearly every part of the production pipeline: scriptwriting, storyboarding, even directorial decisions. Pre-production, production, post – all generated by agents. What would’ve taken months with a traditional team was compressed into weeks. “We personally estimate about 65% of the tasks agencies get paid for currently could be done by AI agents with today’s technology,” said Wes ter Haar, an executive director of S4 Capital on the earnings call for the group’s interim results. It’s the quiet part said out loud. https://lnkd.in/g43YqC32 | 19 comments on LinkedIn

A few #martech and #ecosystem observations on Snowflake's just-released Modern Marketing Data Stack 2026 report: 1️⃣ Cloud data warehouses/lakehouses continue to gain momentum as a universal data layer, driving improved data integration across the tech stack. It isn't a panacea, as there's still work required with good cross-org data design and governance. Still many a Big Ops challenge on top of Big Data. But it's a solid, open foundation to build upon. 2️⃣ Snowflake has one of the best aligned platform ecosystem models in tech today. They make money from their customers consuming compute and storage resources. The more 3rd-party applications a customer integrates with Snowflake, the more resources consumed and billed. Customers only connect and keep using apps that they find valuable. The integration between the app and Snowflake make both more valuable (more data flow = more value). There's near perfect alignment in incentives and rewards between Snowflake, its integration partners, and their mutual customers. The hyperscalers (AWS, GCP, Azure) have similar ecosystem alignment. But it's still a rare pattern among SaaS application platforms, which often rely on marketplace transaction fees, rev share payments, program fees, or indirect measures of value to drive tech partnerships. While those non-consumption partnership models are workable, they often risk "drift" in their alignment. 3️⃣ What's remarkable about Snowflake's Modern Marketing Data Stack illustration below is its relatively good understanding of martech categories and players connected to their platform. Snowflake is not a martech company per se, but this representation of a martech ecosystem is better than I've seen from most pure-play martech platforms. Of course, it's not perfect — I could certainly nitpick logos and categories for hours (it's my hobby). And I assume there's bias around featured partners. But overall, this is a reference example of how a broad infrastructure provider can do excellent partner marketing in a domain. 4️⃣ I love that service providers are represented on this tech ecosystem map too. There's such a powerful interplay between them, but they're rarely shown together in the same context. The blurring boundaries between software and services with AI is only going to make that more of an entangled ecosystem. (Nod to Jay McBain and his 7 partners on average who influence a deal.) Good stuff! | 46 comments on LinkedIn

Login to LinkedIn to keep in touch with people you know, share ideas, and build your career.

🚨 Introducing the AI Apps 50: Startup Edition Ever wondered how startups are spending their money when it comes to AI? Our team at Andreessen Horowitz worked with Mercury to crunch the numbers and rank the top applications by spend. The list + what we learned from it ⬇️ - Horizontal apps have a slight lead over vertical (60% of the list). This includes general assistants (ex. Perplexity) and SIX different meeting support tools (ex. Fyxer AI). But, it also encompasses creative tools and vibe coding tools that are used in roles across orgs. - Vertical apps can augment human labor...or replace it. We're mostly seeing the former - but five companies on the list allow customers to "hire AI" (ex. Crosby Legal, Cognition, 11x). Labor augmenters mostly assist with customer service, sales, and recruiting. - Vibe coding has landed in enterprises. It's not just a prosumer trend! Number three on the list, below OpenAI and Anthropic? Replit. Other listmakers in the category include Lovable and Emergent, while Cursor made the ranks for more technical users. - Products are making the consumer - enterprise jump. 12 cos also appeared in our most recent Consumer AI Top 100 - almost all of which started out B2C and have migrated B2B over time. In fact, 70% of listmakers are available for individual use (no enterprise license needed)! Check out the full report: https://lnkd.in/gmMvfvSv | 15 comments on LinkedIn

Ah, the Hype Cycle. No other curve is both so revered and reviled. When it comes to generative AI's Hype Cycle, especially its application in marketing, there are three things we need to acknowledge: First, 'gen AI' is not a single technology moving along the Hype Cycle. Is it at the peak of inflated expectations? The trough of disillusionment? The slope of enlightenment? The answer is: all of the above. Different gen AI applications and

Cost per token is the WRONG metric to focus on. As businesses integrate generative AI into real workflows, they are facing rising costs due ...

Super insightful map of AI agent tech stack categories by CB Insights with plenty of good examples of vendors/products in each. #martech | 12 comments on LinkedIn