Theory Ventures @Theoryvc

x.com/i/article/2099…

Many concerns this weekend about slowing down AI, some targeting open models. We need to be responsible. But we can't let this slow down, or worse, prohibit open models. It's the wrong risk. Open models have tremendous power to democratize AI and make it more personal. The larger risk I see with open models is right in front of us: in the last week I've read about how many popular platforms quietly send data to foreign jurisdictions, or worse, sell or train on it to gain an advantage. And this is becoming more and more mainstream. Now more than ever open model vendors must act in the user's best interest, not their own: zero data retention or training, and hosting in the user's region vs sending data overseas. This has been our belief and commitment with @ollama. Nobody needs to slow down open models. We need to distribute and run them in a way users can trust.

@jdewitt29, COO of @vercel, joins @ttunguz to walk through something that kept failing a few years ago: running inbound sales with AI. 📅 Oct 1 ⏰ 11am PT 📝 Register here: gatsby.events/theory-venture…

🦙 @ollama is used by 9 million developers and 85% of the Fortune 500, giving co-founder and CEO Jeffrey Morgan (@jmorgan) a unique view into which AI models people are actually using and how that’s changing. Right now, the biggest shift he sees is toward open models, driven by coding agents, falling costs, and capabilities that are rapidly catching up to the frontier labs. On Ollama Cloud, that shift has driven a 150x increase in token usage since the start of the year. In this episode of @LightconePod, Jeff joins @garrytan, @snowmaker, @sdianahu, and @harjtaggar to talk about the future of open models and the story behind Ollama, from two years of searching for the right idea to building one of the most widely used AI developer tools in the world. 00:43 — The Shift to Open Models 03:03 — How AI Agents Are Driving Token Usage 05:31 — Are Open Models Catching Up? 08:26 — What Happens When a New Model Launches 11:31 — Ollama as an Operating System for AI 14:05 — The New Opportunities Above the Model Layer 18:19 — Why 80–90% of Enterprise Tokens Could Be Open 20:57 — The Future Is Local and Cloud 26:40 — Why AI Is Coming Back to Your Computer 28:56 — The Coming Era of Unlimited Tokens 32:30 — Do We Still Need a “God Model”? 33:41 — Open Models and Geopolitics 36:14 — The Origins of Ollama 40:36 — Two Years Lost in the Wilderness 42:39 — The Pivot That Changed Everything 47:02 — How Ollama Found a Business Model 49:43 — Why Second-Time Founders Did YC

More tokens per megawatt -> higher revenue per megawatt -> more profit into the next training run.

Tomasz Tunguz @ttunguz

2 weeks ago

Model companies buy electricity by the megawatt and resell it as cognitive work. Anthropic's revenue has run as high as $50m per megawatt, against $10-15m of compute.

13 16 159 32K 168

The frontier AI market is sorting itself into closed camps, as labs pick partners, cut rivals, & ration access to their strongest models. x.com/i/article/2094…

x.com/Theoryvc/statu…

Theory Ventures @Theoryvc

3 weeks ago

x.com/i/article/2092…

0 0 8 1K 6

When hundreds/thousands of new features ship every month, how do you coordinate project planning? Customer success? Sales enablement? Andy Triedman on what has to change:

x.com/i/article/2092…

Ollama v0.33 is here! You can now easily configure Claude Desktop to seamlessly work with Ollama as a third-party gateway provider. One toggle. Cloud & local models just work👇

🤯 Excited for all the models to come! Ollama is exploding in token usage. Better models, choice of apps/harness, and all private. We are just getting started. Let's continue this record-breaking streak.

My codebase slop theory is there's some threshold past which slop goes exponential

@CaylaSharp2 @BEBischof @ethanyish @AlliumLabs What if we record one on a ferry 🤔

On our latest episode of In Practice @BEBischof sat down with @ethanyish (CEO, @AlliumLabs) to discuss: -> why a blockchain is a Turing machine, not a database -> how one wallet balance = replaying history to 2009 -> 12,000 dbt models, ~30PB, 3% the size of the web -> why AI agents shift leverage to domain experts

0:30 — a data warehouse with batteries included 4:29 — why a wallet balance means summing from genesis 7:45 — the Chief Data Plumber & the cost of curiosity 13:41 — tracing a hacked wallet in one prompt 15:49 — the Omni vs. MotherDuck semantic-layer debate, settled 25:34 — stay for the post-credits scene

Custom software is cheap to spin up, but expensive to keep. @BEBischof on @MTSlive talks about the industry's growing tech debt and the type of platforms needed to reduce it.

Theory Ventures @BEBischof reveals he is still hiring engineering interns because Jevons paradox means AI accelerates work faster than it eliminates it: "I'll mention some comparables. We got Grok Bot yesterday. We got Claude Tag a little bit before that. I'm very excited about these new products. For some types of things, they're incredibly valuable." "The way I tend to use this product is to augment me and to augment others. To give you a barometer of can I replace people with AI yet, I'm still hiring interns. I'm not even saying, oh, I only hire senior staff engineers now. I still hire engineering interns, and they're really helpful and productive and valuable." "The frontier moves. What used to be a real pain in the butt for me, I can now have my coding agent take care of, but that doesn't make any less work for me. It's the standard Jevons paradox here." "To answer very sharply, I think we're a couple years away from knowledge work hiring going down a lot. I think we will continue to embrace Jevons paradox and hire more." @Theoryvc

Watch the full conversation here: youtube.com/watch?v=dGng0Q…

Your wallet was hacked. Now what? Stolen money jumps wallet to wallet, chain to chain. Historically, you'd have to track hops by hand. Now, you can ask your agent.

New episode of In Practice feat. @ethanyish (aka Chief Data Plumber)! Check it out:

Bryan Bischof fka Dr. Donut @BEBischof

a month ago

"It was a bright and sunny Friday afternoon, when a ping came in from the CEO: ' I have a quick question.' " That'll scare your data scientist friends around a campfire! What every data scientist fears is somehow even spookier for the Blockchain. As the 'Bloomberg for Crypto',

4 6 19 3K 8