2X YC founder. Now founder at Vectoral. Working to stop token fraud.

San Francisco, CA
Joined March 2015
Past 30 days' traffic for Vectoral. Content Marketing is a lot of work, and you're playing the long game, but it definitely works if you dedicate the time.
3
27
3,741
Seen in a closed startup forum: "Interested in selling OpenAI credits for much lower price to any startup founders or devs. I got around $2500 worth of OpenAI credits for being a part of YC Startup School."
3
10
491
I’ve seen a lot of people dunking on @thdxr and @opencode, and I needed to say something. Frankly, I don’t think anyone understands how sophisticated these attacks are. These aren't script kiddies doing this after school. These are state actors, with real companies fueling the fire. One of the bigger companies behind distillation has 4 office locations worldwide, and there are a number of companies that are entirely fronts for distillation. You’ve probably seen their ads. Some have raised venture. I’ve personally seen messages claiming they bought over $5 million in inference for distillation. They’ve gone so far as to recruit people close to the model labs for better access, and the dollars flowing through the distillation ecosystem are in the 9 figures. These attacks offer a huge economic incentive. These aren't fly-by-night operations. At Vectoral we've see relays offering affiliate programs, net 30 invoicing for enterprises, and strict SLAs. Together, the top relays get tens of millions of visitors per month. Frankly, they are more sophisticated than most SAAS companies. The sad reality is that things likely only get worse from here. We are at war for global AI hegemony.
32
15
13
212
24,318
Matt Lenhard retweeted
Today, we're launching Sazabi version 1.0. The reality of shipping software as a team in the age of AI is actually crazy. Agents merging their own PRs. Hundreds of deployments per day. A codebase that is constantly changing. Junior developers with 10x or 100x leverage. 😵‍💫 I spent the last year building a safety harness for teams that ship with AI. It's called @sazabi. The premise is simple: 1. You send us logs. 2. We detect anomalies, investigate them, and send you back rich, actionable alerts about real problems in your app. Sazabi helps you find issues before your users notice. No monitors or noisy alerts. Just pure signal. And today we're making it available for anyone to use.
123
75
83
496
641,121
What makes token fraud even more dangerous is that most companies don't know it's happening. If attackers are only taking a couple thousand dollars a day spread across enough keys, it may never get flagged. It only surfaces when they push into six or seven figures. By then you've already lost a lot.
1
1
10
306
The token relay market has built all of the same commercial infrastructure as a legitimate SaaS company. - Price comparison sites - Affiliate programs - Customer support WeChat groups - Model verification tools - Enterprise procurement pages - Net-30 invoicing - Data Processing Agreements
3
1
1
14
690
Our 2 preprints just dropped. We're doing neuro drug discovery backwards. First, we use ML to identify brain activity patterns that translate from mice to humans. Then use CircuitATLAS to find molecular targets that can change them. Two halves of the same discovery engine. Papers below 👇
1
6
12
752
AI has made coding more fun but way less fulfilling
13
442
We're hiring across a bunch of roles at Vectoral You'll be at the forefront of stopping some of the worst things in AI. Willing to go above and beyond to build a team of exceptional people. If you think that's you, DM me or apply here: jobs.ashbyhq.com/vectoral
40
14
2
429
28,179
I've become convinced client-side bot detection will never work. CAPTCHAs are DOA.
2
1
21
271,481
Insane how much of a difference a great hire can make. By far the highest leverage thing I've done at Vectoral so far.
1
1
27
965
I made a coding agent that’s entirely free no credit card, api key, or subscription, forever we use an efficient harness and open source models so you can use frontier quality models all day there’s a reason the most iconic consumer companies are all free :)
57
34
19
267
33,286
Someone is selling 100K OpenAI and Anthropic keys right now. Crazy times. A bunch of people are about to wake up to huge bills.
3
2
16
1,142
Commodity markets have a concept of provenance, where there is proof of origin, grade, and chain of custody. We really need something similar for the inference market.
be careful in your pursuit of cheap tokens we've come across so many sketchy things that are out there with tons of usage remember you're hooking up your harness to these inference providers. they can and do crazy things like send fake tool calls to steal your info they also sell entire traces to third parties who resell again. a single key in the trace is enough to screw you if you cannot explain how someone can offer cheap tokens you should avoid it
5
1
20
2,047
Have a feeling the ATO attacks are about to get a lot worse. Seeing a ton of posts like this.
2
5
400
I don't understand the whole AI agent persona trend. More marketing than substance, unless I'm missing something.
7
13
1,187
Why do all of the inference attacks spin up thousands of concurrent requests? Are the labs sampling on the safety identifier? Or maybe bans are asynchronous, so attackers try to churn and burn before they get caught.
2
3
715
Anthropic is doing everything they can to make sure that the average person hates AI
4
15
486
The pace and quality at which the @mastra team ships needs to be studied. Pumped to try out factories.
Software engineering is becoming a hierarchy of loops. So today we're launching Mastra Factory: a system for agents to take software from issue into production. Get started with 𝚗𝚙𝚖 𝚌𝚛𝚎𝚊𝚝𝚎 𝚏𝚊𝚌𝚝𝚘𝚛𝚢:
1
4
18
2,705
The Token Relay / Reseller market reminds me a lot of the residential proxy market. If you are looking to hide who you are or get around restrictions, you use a proxy. For inference, you use a relay.
1
13
700
You can now white-label your own token fraud business
3
1
14
745