Bradley Reynolds

Nooks just launched AI Sequencing The company built on parallel dialing just admitted they need the layer which Outreach has been selling for years …and we all watched Outreach’s CEO step down, survive 5 rounds of layoffs, and their $4.4B valuation dwindle Meanwhile TitanX raised $27M and acquired Frontspin to do the exact opposite Precision over volume - one perfect dial instead of burning through your entire TAM at 2-4% connect rates Mass outreach is dying and the market is picking sides Here’s what Nooks may not tell you: their dialer runs on Twilio infrastructure Which means the phone numbers your reps are calling from aren’t even theirs - they’re provisioned through a third-party telephony provider that competitors say causes “lag time and stability challenges” Fun fact: Nooks actually recommends you register their numbers with freecallregistry*com to avoid getting spam flagged …tells you everything about the quality of numbers they’re providing Parallel dialing works when you have fresh data and buyers answer unknown numbers - but when your reps are calling from numbers that might get flagged as spam, half the data theyre use was scraped months, even years ago, and most cell phone providers enable unknown number call screening You’re just reaching failure faster Nooks built their business on “dial 5 numbers at once and skip the voicemails” but when connect rates keep dropping and customers ask “why are we paying to leave more voicemails faster from numbers that get spam flagged?” You need a new wedge - so now they’re adding sequencing, signals, and AI personalization Sound familiar? Nooks is following the same playbook as Outreach - bolt on features and hope customers don't notice volume based outbound is solving yesterday's problem The companies winning right now aren’t competing on how many prospects you can reach in an hour - they’re competing on accuracy of prediction before you reach out at all TitanX: “who will answer” NextLM: “who’s showing buying signals right now” Both built on the same thesis: your reps don’t need more volume, they need better precision and perfect timing There are 2 opposing forces in the market right now and 1 is clearly winning When even the parallel dialing company admits you need signals and mass sequencing to make the outreach work, maybe it’s time to stop paying for platforms that measure success by number of outputs per day Your team is only as good as their list - and your list is only as good as knowing who to reach out to, and when​​​​​​​​​​​​​​​​

69

46 Comments

As a VC or founder, you should be focusing on consumer. GPT, Claude, Lovable, Cursor all have one thing in common. Their growth was led by the individual consumer. AI unlocks massive value for individual consumers, in a sense that it disintermediates expertise and intelligence, and changes content interaction and buying behavior. Still, the focus of YC and most VC’s is still on ‘vertical AI’ which doesn’t require a massive leap of logic from the Vertical SaaS gold rush of the past decade. I get that vertical AI will unlock massive value and may be an easier wedge, but I’m yet to see anyone even remotely interested in the possibilities unlocked for consumers. AI has unlocked new content interaction possibilities, new ways to manage personal knowledge, and new behavioral patterns are already emerging. Consumer has been out of vogue for the past 3-4 years now, and now with consumer purchasing power increasing, right now may be the best time to build / grow a business. 3-5 years from now when consumer is the flavor of the month, it may be too late to dislodge the ones who’ve built on conviction and clarity today. Side note: I understand that the chat interface is still the preferred UI, but does no one else find it just a bit clunky / inconvenient? Some food for thought. 😏

41

4 Comments

Just got a pitch email from a company selling AI based diligence to VC firms. They saw an announcement for one of our VITALIZE Angels companies called Airia, which is an enterprise security platform. The DD report they sent me (using Airia as a test case) did a market analysis on home fragrances, candles, and essential oils. LOL. The majority of what I see in AI diligence is just not good. Some of it gets halfway there, but a real person MUST look at this stuff before saying it's accurate. Anyone seeing any advanced diligence / workflow tools that get the AI right 99% of the time?

41

27 Comments

Fortune 500 CMOs: Google just dumped your playbook. Google’s shift to AI Reviews means the billions you’ve invested in SEO and SEM are at risk of becoming stranded assets. Here’s the reality: LLMs (ChatGPT, Gemini, Claude) compress your content into answers without guaranteed attribution. The funnel is no longer clicks and blue links - it’s AI-driven discovery. If an AI model defaults to recommending your competitor, you’ve already lost. The AIVO Standard™ (AI Visibility Optimization) is the only framework designed to protect Fortune 500 brands in this new paradigm: ✅ Enforces entity integrity so AI models describe your brand correctly ✅ Secures visibility across ChatGPT, Gemini, Claude, and Perplexity ✅ Provides auditable, cross-team governance for marketing, data, and comms The options for CMOs are clear: 1. Do nothing → lose visibility and brand equity 2. Patch with “AI hacks” → short-lived, inconsistent 3. Adopt the AIVO Standard™ → build durable visibility and resilience Google has already walked away from traditional search. The question is whether your brand will still be visible in the AI era. 👉 Fortune 500 CMOs: which option are you choosing?

3

1 Comment

"Using" AI is, I'm afraid, tablestakes. We've been doing that for some time - even if we didn't know it. Building with AI is a different thing entirely. I like to think of it from a leverage perspective - AI is a multi-modal simple machine. It makes big work smaller and less difficult. This also touches on the whole "the bots are coming for you" take we often hear. I don't think AI "wants" to replace humans, but if we don't embrace it and make a core part of our workflow, someone who is doing that may certainly pose a risk.

3

🤖 Anthropic just gave Claude the ability to take over your computer. Open files. Click buttons. Crawl the web. Run code. In the same breath as the product launch, Anthropic warns users not to run it near sensitive data and disables certain apps by default. That's the company telling us the guardrails aren't ready for the power they just shipped. This is the core tension in agentic AI right now — the capability is outrunning the safety infrastructure around it. An agent that can click, scroll, and execute on your behalf can also be hijacked, manipulated, or simply wrong in what it gets up to while you're not watching. The competitive pressure from open-source frameworks like OpenClaw is clearly compressing timelines. Labs are shipping autonomy before they've nailed the governance layer. We're in the phase where the car exists and the seatbelts are optional. That gap between capability and governance is exactly where I'm focusing right now. Industrial applications simply can't operate in a no-seatbelt situation — and the companies building that safety infrastructure are one of my hottest investment thematics.

52

7 Comments

Back in the late 1990s, when I was an associate in a CVC in the U.S. named Greenhouse Fund, one of the most fascinating case studies we analyzed was the “browser wars.” Netscape Navigator had captured early market dominance through innovation & ease of use, while Microsoft’s Internet Explorer leveraged Windows distribution to overtake it—an early and powerful example of platform advantage. For VCs, it was a vivid reminder that in technology, product innovation alone rarely wins without strategic control of distribution and ecosystem lock-in. When the second wave arrived in the 2000s, we watched Mozilla’s open-source movement challenge incumbents on transparency and standards, and Google Chrome redefine the browser as a true operating system for the web. Chrome’s speed and integration with Google’s services made it the new center of gravity for digital life and a model for how to scale globally through relentless user-centric design. Today’s emerging “browser war,” now between Chrome (Google), ChatGPT Atlas(OpenAI), and newer entrants like Strawberry, feels strikingly familiar but more intelligent. The competition is no longer about rendering speed or interface design. It’s about who can best integrate AI into the browsing experience. For an investor, it evokes déjà vu: the next platform shift is already unfolding, and the real battle is for context, not clicks. Only once in a generation will we see such browser wars again, and the question is who wins this one? thoughts?

53

10 Comments

The AI race is moving FAST! xAI is finalizing a $9.3bn capital raise, $5B in high-yield debt and $4.3B in equity to finance data center expansion. Institutional demand is strong, with TPG Angelo Gordon anchoring $1B of the debt. This raise follows xAI’s March merger with X Corp at a $113B valuation and a secondary share sale to validate that pricing.

13

1 Comment

Found this interesting article "Do AI-enabled companies need fewer people?" by Laurie Voss within the Venture Capital topic on AI Colorado. AI-Native vs Traditional SaaS: AI-native startups operate with roughly 60% fewer people at $10M ARR (15-20 vs 50-70), generate 6x higher revenue per employee ($3.48M vs $580K), and raise 30% larger Series A rounds ($51.9M vs $39.9M). Layoffs Continue: Over 126,000 workers were cut at US tech companies in 2025, with the pace carrying into 2026. Tech job growth has flatlined since 2023 in what the author calls another K-shaped graph. Interesting times ahead.

11

2 Comments

The AI agency market is not saturated. You are just looking at the same 200 people online. Nick Saraev , Liam Ottley , Nate Herkelman . Aaron shared a stat today that should make a lot of founders rethink the entire market: 1.7M US businesses need AI implementation. Around 1,500 AI agencies exist. That is 1,100+ potential clients per agency before this market even gets close to crowded. So no. You are not late. You are early enough that the biggest advantage still is not capability. It is visibility. Because most founders are confusing internet noise with market saturation. They open LinkedIn. They see AI agencies everywhere. Automations. Agents. Systems. Workflows. And they assume the market is full. It is not full. It is just loud in your feed. The real market is still wide open. Because the actual buyers are not the people posting about AI every day. They are the business owners quietly falling behind. A dentist in Ohio. A logistics company in Texas. A law firm in Miami. A 20-person team that knows they need to “use AI” but has no clue who to trust. They are not in your comments. They are not reading TechCrunch. They just know one thing: Something is changing. And they do not want to be late. That is the opportunity. A massive market with confused demand and very little clear supply. And that is why this stat matters. The bottleneck is not demand. It is not talent. It is discoverability. The businesses who need help do not know who to call. And in markets like that, the winners are not just the best operators. They are the most visible trusted option. The name they have seen before. The founder who already feels credible before the call. That is how early markets get captured. Through repeated exposure. Through trust at scale. Through being known before everyone else wakes up. And this window will not stay open long. Right now, even average agencies can win because the floor is low. But that changes fast. More agencies. More noise. More identical offers. And once that happens, visibility is no longer an advantage. It is the minimum requirement. That is the shift most people miss. They think they can build distribution later. Later usually means after attention is gone. After the market has already chosen familiar names. That is why the market is not crowded. The visible part of the market is. And the visible part is where buyers decide. That is exactly why we built Vision Reach AI Because most agency founders are stuck in delivery. Client work. Operations. Referrals. They are serving the clients they have while staying invisible to the ones who could buy next. So we handle the distribution side: → LinkedIn ghostwriting → Newsletters → EECs → X growth Built to help agency founders become the obvious name in their category before the rest of the market catches up. You are not late. You are just not visible enough yet.

10

1 Comment

🔥 What. A. Day. Nearly 200 leaders from 7 states came together in Louisville for the first-ever Vsimple AI Summit & the energy was electric! The aim was not to be a traditional "AI conference", but for it to be educational, fun, inspirational, & a turning point: ->Leaders walked out on fire to start their operational transformations & harness the power of AI in their organizations/communities! ->Conversations shifted from “someday” to “what do we do in the next 90 days?” ->The message was clear: AI isn’t the future, it’s the present, & the leaders who move now will define the future of their industries. The theme of the Summit said it all, “Where the Bold Gather, the Future Follows” & the people in the room proved themselves BOLD! Louisville Business First captured the momentum, writing about how Louisville is stepping into a new role as a hub for AI & how the conference was "a combination of both an academic exercise and a pep rally", which is exactly what we were going for! 👉 Article below To everyone who joined us -> THANK YOU for believing in this vision, for showing up ready to lead, and for proving that the bold truly did gather here in Louisville this week. The AI revolution has started, the question every leader must answer: Will you be a follower… or will you lead? #AISummit #FutureOfWork #AILeadership #Louisville #UncommonEverything #VsimpleLife

91

15 Comments

While everyone was debating whether AI would replace consultants, OpenAI just became one. Their new consulting arm charges a minimum of $10 million per engagement, embedding Forward-Deployed Engineers directly into client organizations to build custom GPT-4o implementations. This isn't just another enterprise sales play. It's vertical integration disguised as consulting. Despite OpenAI reaching $10 billion in annual revenue and serving over 500 million weekly users, they're still burning through approximately $5 billion annually. API subscriptions alone aren't cutting it. If OpenAI, which started the ChatGPT revolution, needs to move into services, what does that mean for smaller AI startups? Google, Anthropic, Meta - they're all building comparable models. The differentiation isn't in the core technology anymore. It's in implementation, integration, and delivering measurable results. So what's the play for everyone else? 𝗕𝘂𝗶𝗹𝗱 𝘁𝗵𝗲 𝘁𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆 → 𝗢𝘄𝗻 𝘁𝗵𝗲 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 → 𝗗𝗲𝗹𝗶𝘃𝗲𝗿 𝘁𝗵𝗲 𝗼𝘂𝘁𝗰𝗼𝗺𝗲. OpenAI is shifting from the first to the second. The real money is in the third. The future belongs to vertical solutions that understand specific industries deeply enough to deliver turnkey outcomes. Not APIs. Not platforms. 𝗥𝗲𝘀𝘂𝗹𝘁𝘀. This is why I'm betting on micro-SaaS in specialized verticals. When AI becomes commoditized infrastructure, domain expertise and implementation capability become crucial differentiators. Pure technology plays are getting harder and harder to defend. APIs are becoming commodities. Implementation is the new IP. I share more musings on my substack >> https://t2m.io/4aDyNCpu For the love of the game 🏴‍☠️⚡️ #ai #futurism #chatgpt #openai 𝘐𝘮𝘢𝘨𝘦 𝘊𝘳𝘦𝘥𝘪𝘵: 𝘚𝘪𝘭𝘪𝘤𝘰𝘯 𝘝𝘢𝘭𝘦𝘳𝘪𝘦 𝘉𝘦𝘳𝘵𝘦𝘭𝘦 

1,393

156 Comments

A very clearly articulated tech stack architecture for enterprise AI with room for innovation at every layer. Thank you Arvind Jain For enterprise AI founders this is a good reference post on which layer you plan to dominate For vertical AI apps this provides a great framework on how you can fit into your enterprise clients' tech. stack/ ecosystem Vik Ghai Amar Chokhawala Shivani Govil

17

2 Comments

This stat about Claude Code is making the rounds: $200/month Claude Code plan can consume $5,000 of compute. People are trying to make this sound alarming. It probably isn’t. Anthropic hasn’t officially released the data, so can’t really verify it. But what’s likely happening is that a small number of power users (running agents, long chains, autonomous workflows) are burning huge amounts of compute. Meanwhile the much larger base of developers use it for simple things like autocomplete, quick prompts, and small refactors. Not much compute. So the net is that it’s nowhere close to 25x subsidy that people think it is. It’s more like an insurance company where everyone pays the same monthly premium and a large percentage of people never use it. The insurance company ends up doing just fine. If someone pays $200 and consumes $5,000 of compute, the gap is being funded by model providers and their investors. They’re temporarily paying the bill to embed these models into developer workflows. The bet is to eat the compute cost today and own the developer platform tomorrow.

60

39 Comments

Think all the value accrues to the models? Look at Anthropic's pricing. For my usage they're subsidizing tokens ~90% through the Claude app vs the API with OpenClaw. They clearly see massive value in owning the harness, not just being the model underneath.

18

5 Comments

AI is no longer just a tool. It’s rapidly becoming the backbone of industries. For years, we’ve debated AI’s capabilities: models, algorithms, and parameters. But at this year's GTC, the conversation has shifted. It’s no longer about “Can AI do this?” but “How will AI disrupt and reshape entire industries?” With 39,000 attendees from engineers, entrepreneurs, investors, and professionals from all fields, the big question is clear: Will AI redefine the workforce and productivity in the next decade? I’m certain it will. While many jobs will remain, how we work will undergo a fundamental change. This is why Jensen Huang has been saying, “AI is the new electricity.” Just as electricity reshaped the industrial era, AI will reshape the digital era. What’s truly striking is how quickly the industry is evolving. We're not waiting for one singular breakthrough. AI is already transforming the landscape in ways we’re only beginning to understand. #AI #DigitalShift #Productivity #FutureOfWork #Innovation #TechRevolution

1

Hot take on AI's value as a product in our daily lives. 1. For most, AI is a nice-to-have, not a must-have It is impressive how AI creates content, generates spreadsheets, vibe-codes, fills CRMs and insurance claims, triages patients, classifies accounting transactions or mimicks human interaction. The wow effect disappears when you start using these capabilities intensively as part of your job. It is somewhat helpful, but not a dramatical improvement to your productivity, lest of all a replacement for a single human. For companies, productivity gains will be either wasted in longer coffee breaks (as has been the case with most software efficiencies in the past 20 years) or obtained via AI-disguised offshoring. 2. The intrinsic limitations of its foundations will make it hard to be really impactful in our lives AI is a set of algorithms involving statistical classification of data and prediction/generation based on available datasets. You should instinctively mistrust anyone pronouncing the "agentic" word. AI does not really "learn" or "think". No Colossus data center and smartly coded software running on it can ever get close to emulating the 86 billion neurons and 100 trillion synapses the human brain has. A self-adjusting or recursive algorithm can be described in metaphorical terms to make it more understandable and marketable, but a metaphor is not an identity. There will be improved, more efficient and adaptive AI algorithms for sure. We all have seen the improvement in AI videos in the past few years. But they will still be algorithms and unable to create, strategize, or read real-life scenarios and nuances. 3. It is not feasible to get rid of AI's dirty secret, the third-world training workers One million people work as AI trainers, tagging images, providing voiceovers and interacting with chatbots, making $1-10 / hour. This is not a transitory task that will be completed once all models are fully trained. As datasets evolve and models are expected to produce new outputs, further human training is necessary. AI cannot train AI without spiraling into poor human-grade outputs. I asked our chatbot friend Claude to run some math for me. Turns out 2 billion hours per year in productivity gains are directly offset by an equivalent number of annotator/trainer hours. 4. The real cake that core AI companies are chasing: the Internet itself i.e. marketing and content If all of the above is true, why are VCs pouring billions into core AI companies? It is a high stakes game, actually. Google has a near monopoly on search, with a market cap of $3.8 trillion and $400bn in revenue (of which 75% content/advertising). MarTech is $1tn in revenue. Social networks is $400bn. All of these can be won with AI. Content generation, search, advertising. Attracting search users to use chatbots instead. A $100bn investment in OpenAI would surely be a homerun if it fundamentally changes the world, but it's also ok if it only changes the Internet.

17

4 Comments

🚨 New SOTA Alert: This Open-Source Chinese Model Just Outperformed GPT-5.1 and Claude 4.5. The landscape of Code LLMs just shifted. While we were focused on the usual giants, the IQuest Coder Team dropped a bombshell with the IQuest-Coder-V1 series. This isn't just another fine-tune; it’s a fundamental architectural evolution in how we train models for software engineering. I've dived into the technical report, and here is what you need to know about the new King of Code. 👇 https://iquestcoder.ai/ 💡 The "Code-Flow" Revolution Most models are trained on static code. IQuest introduces a "Code-Flow" multi-stage paradigm that mimics the dynamic evolution of software logic . * Pre-training: Transitions from general data to high-quality code annealing. * Mid-training: A dual-phase injection of reasoning and agentic trajectories at 32k and 128k context lengths . * Bifurcated Post-training: This is the game-changer. They split the training into two specialized paths : * 🧠 Thinking Path: Uses Reinforcement Learning (RL) for deep reasoning and error recovery. * 🛠️ Instruct Path: Optimized for general assistance and tool use. 🚀 The "Loop" Architecture (Efficiency Hack) They introduced the IQuest-Coder-V1-Loop. It uses a recurrent mechanism to balance model capacity with deployment footprint . Essentially, it allows for iterative computation over complex code segments without the massive compute tax of traditional dense models. 📊 The Numbers (Receipts) According to their technical report, IQuest-Coder-V1-40B achieves State-of-the-Art (SOTA) status: * SWE-Bench Verified: Scores 76.2, beating proprietary leaders . * Code Reasoning (CRUXEval): The "Thinking" model dominates with 98.5% accuracy on Output-COT, surpassing Gemini 3 Pro and Claude Opus 4.5 . * Head-to-Head: It shows competitive or superior performance against GPT-5.1 and Claude 4.5 Sonnet across HumanEval+ and MBPP+ benchmarks . 📦 Versions & Availability The team has released the complete white-box chain of checkpoints . You can download them now (likely on Hugging Face given the open-source standard): * Sizes: 7B, 14B, 40B . * Variants: Base, Instruct, Thinking, and Loop . ⚠️ Limitations While the performance is massive, the 40B dense model requires significant hardware resources for inference. The Loop variant is designed to mitigate this , but for maximum reasoning capability (The "Thinking" path), you are looking at a heavy compute cost. The Verdict? The gap between open-weights and proprietary models isn't just closing; in coding, it might have just disappeared. Have you tested IQuest yet? Let me know in the comments. 👇 #LLM #ArtificialIntelligence #Coding #MachineLearning #OpenSource #AI #IQuestCoder #TechNews #Innovation

8

3 Comments