为 Instinct 辩护：个人 AI 新实验室为何可能值千亿美元

Ken Wattana @KenWattana原推发布于 2026.09.28

在 X 上打开原推

A lot of talk about why Instinct will fail, is a fad, bubble, the next Clubhouse, etc. Big valuations in short periods of time get people going. So let’s be optimists and think about why they can win and the VCs are betting on them in size, rapidly. Abundance. The way I think about it - Instinct is the neolab for personal AI. They will require massive capital expenditure to push the limits of what's possible. If it works, the outcome will be multiples of the price they pay. If not, they still can be acquired for a good exit given the caliber of the team and what they've achieved (and will be able to achieve - ex. if given unlimited compute like Cursor). The pressure lies on the growth investors who have to underwrite the risk at this stage. Why I think they Instinct can be a category defining company (ex - $100B+) 1) they brought magic back to consumer AI. Instinct was really the first personal agent to make things work smoothly across all the big categories. Booking, shopping, negotiating, planning, scheduling, communicating. Success rates across the board are high and it feels like magic. The less human intervention needed, the better. Pair it with some clever onboarding patterns, it genuinely feels much better than previous generations of personal agents. This was hard, almost impossible to do until now - so it’s a big differentiator. Any failures really shatter the ideal of the personal agent experience so it’s probably a top metric that they look at. Meta has caught up with their own browser use with Muse, but I think Instinct will continue to push the frontier in regards to agentic actions, whether that is browser use or integrating APIs and protocols. Product remains a significant differentiator as models get commoditized. People don't care about the engine as much as they do the car and the end experience. 2) team. Noah was early at Sierra. Did research on computer work. And most of all, shipped something that flipped the whole category on its head. Beyond credentials and reputation, that's what stands out to me. The personal agent space was filled with products that were really more expressions of "wouldn't it be cool if it could do xyz". But none of them could actually do it. When you find a founder that has a clear and unique vision and can execute and define the category, that's someone you want to double down on. If you don't think consumer AI will be zero sum, Instinct is the team you want to bet on. Working at Instinct won't be for everyone given how stratospheric the valuation went, but I do think this will still attract talent for people who want to go head to head with the big dogs, build something interesting in an era of clones, and move fast with a smaller team. 3) capital, yes it matters. Investors moved fast to kingmake Instinct. And why not, it's the first personal assistant to have that lightbulb moment. And when you're taking on Meta/Apple/Google/SpaceX/OpenAI, you need as much money as you can get. A lot will go to compute in addition to growing the team. With capital also comes connections and I imagine a big deal and partnership is in the works with some one like a Microsoft who can offer a wealth of compute without being directly competitive in the way other compute providers would be. Instinct is fighting a battle on all fronts - in product competition from the biggest companies in the world, in a scarce compute market, and in a war for talent. This is not an operation that you hope to get profitable fast and never raise again. Fundraising is a significant variable in the equation. 4) market strategy - there is always competition. Why even go head to head with Meta at all? A lot of investors will sit this one out just because of that, but I also think a lot are willing to make a contrarian bet in the space. The consumer AI market is massive and will only continue to grow. There's room for several winners. Meta can’t and won't win it all, despite how all-encompassing they may feel when the PR and billion user distribution machine is firing. If you’re an investor not afraid to take risk and fund something competitive, this is the best team to do it. No team captured user interest or had a product like Instinct in the personal assistant space the past three years. Conviction made their name this cycle by reasoning that AI is not zero sum and only for hyperscalers or 1-2 labs. There will be big businesses built and opportunities for sharp teams. That said, their thesis has already been proven right - it's on the late stage investors to continue to believe and have the patience that there can be outcomes outside of the labs and tech giants. But if you don’t take these swings, why are you a VC? It’s not for everyone or every fund at this stage for sure. One thing I notice is that big labs can’t get skinny, too much to do, too many people to build for. It's hard to build unique products for them now. Similar to being a large company. It's also why Muse's success is so interesting and unique. And why Grok Bot (mainly driven by the Cursor team) shows the impact of an empowered team within a bigger organization. OpenAI, Anthropic, are also dealing with massive challenges every day and revenue was prioritized for IPO. They're playing catch up here. I don't think Anthropic really wants to play at all. I expect something from OpenAI at Dev Day, but their consumer product has been neglected as they shifted focus to Codex. 5) shipping fast - yes this can make a difference. It seems like there's a new feature almost every week. Instinct makes a clear effort to constantly push the ball on this front, particularly building products around building network effects and a sticky experience. Pace of shipping will continue to be critical with Meta and others in the game. New features, unique products, things unique to the Instinct network. There's a million things they can do - choosing the right ones will be important. Some things will hit, others will be forgotten. Many will be copied by Meta like they did with Snap. But this is how you explore new user behavior. 6) roadmap - ambition. Instinct has already signaled its intent to own compute, chips, and models. Aligning all these elements makes for a better product. Controlling the supply chain end-to-end and not being beholden to anyone is what they need to go head to head with Muse and Meta's unlimited capital and compute. They need to think big and it's why they're so aggressive on the capital front. If you want to scale to hundreds of millions of users, you need to take big swings and think outside the box a bit. 7) growth, yes this is the biggest question mark. AI made everyone forget how consumer social works, it’s not all about optimizing for revenue. A lot of people get hung up on the fact they're not monetizing yet. There will be a time for that but it's not right now. If they can add users at a rapid rate, it's just as good but they need to unlock viral growth to keep momentum going. Beyond organic and referral growth, partnerships will be big here and as a first mover there are a lots of levers to pull, not just organic. Commerce partnerships, integrations with technology partners, a distribution deal with a tech giant that doesn't have a good assistant (Microsoft/Apple/Amazon), etc. I wonder if some of the classic deals Facebook made back in the day to increase their reach with publisher and others find a new form here. Exclusive partnerships might end up being a thing. Curious to see if they try to tap influencers and creators at all. Will they focus on being luxury software and serving their power users? Or go down a bit to broaden access. I don't have any answers here - I expect they will hire the best in the business to try and solve it. How will they market and push in comparison to Meta? VCs sharing the product on Twitter will only take it so far, how they break out into the wider consciousness will be key. --- Personally, I think rolling over and conceding what will be a trillion dollar category just because Meta/Google/OpenAI has made a competing product is one of the most boring ways to view the world. VCs exist to fund disruptors and competitors and this is exactly what Instinct is. Big bets, big outcomes. Yes, they will need to play catch up in many areas, namely compute, user base (advertising to a few billion users for free is nice), team, and relevance, but every company started from zero too. I'll always root for the underdog.