
just wrapped up our 7th batch of @conviction Embed 💜 we’re lucky to have spent the last two months with some amazing companies (8 listed below, with a few more still in stealth). here's what they are building 🧵
✓Investor + Engineer @ @Conviction

just wrapped up our 7th batch of @conviction Embed 💜 we’re lucky to have spent the last two months with some amazing companies (8 listed below, with a few more still in stealth). here's what they are building 🧵

@mvernal and I, and all of us here @conviction are excited to announce our investment in Mt. Thor! There should be an Apple neocloud. As agents work alongside us, they should use the same tools we have access to — for most of us, that means they should have a Mac. At a token level, Apple silicon should clearly be used for inference as well. Apple’s silicon engineers have built the best-in-class product for power-constrained on-device compute, and open-soruce models optimized for it will be power and cost-optimal. Peter is tactical, dilligent, and above all else, impatient. When I called him Sunday night at 8 pm in the middle of dilligence, he was still at the data center, manually racking machines and checking them (where he’d be for at least the next copule hours). Everything at Mt. Thor moves a mile a minute, but chronicled dilligently in comprehensive weekly updates, delivered without fail every Saturday, that unflinchingly describe the wins and losses of the past week. It’s an pleasure to work with someone with so much urgency, yet also disciplined. Excited to be on the journey with our friends @crv @chadbyers @offlineventures. To more compute!

Keenable 1.0 is live Two months of learning from agent queries made Keenable the best web search API for agents: #1 on Artificial Analysis and NEEDLE. 1.0 adds 100M knowledge cards and dynamic snippets that cut tokens by 50% on research tasks. keenable.ai/blog/keenable-1.…


Keenable takes #1 on the @ArtificialAnlys Search Index. Less than three months after opening to the public, and three weeks after taking #1 on the Vals Web Search Index. We built Keenable to learn from every agent query and it’s learning fast. artificialanalysis.ai/agents…



I think the "Europe is hard," "early stage is hard," etc. takes are all pretty basic. I think a better frame for this is basically Christensen's Modularity Theory, i.e., most industries start off vertically integrated and then modularize over time for both scale and efficiency. The distinctions between seed, venture and growth are just modularization applied to venture. The problem with modularity theory is that the late-cycle modularization is also the root of the Innovator's dilemma. It's the same forces that cause Amazon, et al to run a consumer service at breakeven and make all of their operating profit via advertising. And then be trapped when personal agents show up. You have exquisitely optimized your business but made it hard to evolve when the operating regime changes. The right answer – which is *really* hard to do – is to throw away the modularization you have created, re-vertically integrate and attack the new regime in a simplified, coherent way. And then in 10 years the cycle repeats and you figure out the new way to modularize and build an exquisite machine for the current regime. Staying true to your roots is great and all, but an equally interesting question is to ask whether you are exquisitely optimized for the old regime or prepared to hill-climb the new regime.

Arjun (@QuantumArjun) eventually figured it out. Hacker 80 was a mis-transcription of "Pat Grady" (say "Hacker 80" really fast). @gradypb – I still think you should try to take over the @hacker80 username.

For a good day-and-a-half, we were incredibly confused. Who the heck is Hacker 80? @sequoia had led the round and there was no Hacker 80 on the cap table.

Fun story: Back in September, @EricNewcomer published a VC Sentiment Report. Question 5a was "What was the most competitive early-stage round over the past 3 months?" One of the answers was @trajectorylabs with this quote.

Congrats to the team on the launch and excited to be re-united with @gradypb and @laurenmhreeder on this as well. And thank you to @gefkovicz and @benmtaft for the initial introduction!

FDE is all the rage atm – and there is definitely a time-and-place for FDE – but FDE can sometimes become a substitute for the hard work of distilling a complex set of topics and skills into a single, re-usable platform. The @trajectorylabs has done that hard work and are now launching that platform today.

The @trajectorylabs team has been one of the best teams I've worked with in my career. The entire team is so smart, hard-working, fast-moving, earnest and customer-obsessed. We love working with them (and you will too!).

Oh no oh no oh no. :( ST:TNG was so seminal to my childhood and Patrick Stewart was the ultimate captain. I’ve had 72hr of air travel in the past week and used it to finally watch Picard Seasons 1 + 2. I literally messaged my brother about it five hours ago. :( Thoughts to his friends, family and all his fans out there. :(

We raised a $150M Series C at a $3B valuation, led by @sequoia. Our third round in 12 months, bringing our total funding to $277M. Our team has grown 10x in the last 12 months. We now serve Fox Corporation, WPP, Live Nation, Lucid Motors, and Ramp, and many of them are fully replacing legacy platforms with @getserval. AI coding tools have transformed how engineers work. Serval brings that same power to every team, so anyone can build governed automations across the systems they already use. We’re hiring 500+ people over the next year and expanding in San Francisco, NYC, London, Austin, and beyond. If you’re down to take on legacy enterprise systems and transform how teams work, my cofounder @AlexAtServal and I would love to meet you. It’s Serval now.