Pinned Tweet
We've grown 3x since April, crossed X M in ARR, and are hiring more folks to join our small lean team
The only requirement for us is you are AI native, can spend more than $20k a month in tokens, and can figure things out in an extremely fast paced environment
We don't have a s, tickets, etc., we don't have a CRM. We operate on Slack and we operate with super high context
Engineering: you have experience building recommendation systems, bidding platforms
Sales: you have experience selling into enterprises while balancing super short sales cycles
Support: you have experience doing support with explicit SLAs and response time thresholds
You will have meaningful equity ownership, and we will work together to scale Lapis from X M ARR to XX M ARR
The only requirement for us is you are AI native, can spend more than $20k a month in tokens, and can figure things out in an extremely fast paced environment
We don't have a s, tickets, etc., we don't have a CRM. We operate on Slack and we operate with super high context
Engineering: you have experience building recommendation systems, bidding platforms
Sales: you have experience selling into enterprises while balancing super short sales cycles
Support: you have experience doing support with explicit SLAs and response time thresholds
You will have meaningful equity ownership, and we will work together to scale Lapis from X M ARR to XX M ARR
19
6
124
72,879
The datavendor team is one of the fastest growing companies I've personally interacted with. They are bringing more transparency to the extremely grey data market, and if you're a co that's exploring selling data in any capacity, you should check them out
Very proud to have them as a Lapis customer from Day 1, congrats team!
Very proud to have them as a Lapis customer from Day 1, congrats team!
We're announcing the public beta of datavendor.ai
We spent the last two years building the best solution for people buying and selling data, whether you're a researcher in a lab or a company trying to sell data.
Data is the fundamental bottleneck for AI progress, and infra that solves it.
4
3
34
2,969
Railway is compromised, rotate all API keys as soon as possible
Deployments are down so don’t wait to redeploy
Attackers are draining keys in real time, we didn’t lose much but can confirm this is still an active attack
Stay safe friends
Deployments are down so don’t wait to redeploy
Attackers are draining keys in real time, we didn’t lose much but can confirm this is still an active attack
Stay safe friends
122
497
4,746
563,086
The more interesting thing with the roadtser is its thrust would be a much more practical car on Mars than on Earth. With one third the gravity the range would be much larger
Given Elon thinks much further ahead than anyone else, won’t be surprised if he already has a fleet of vehicles for Mars in mind
Given Elon thinks much further ahead than anyone else, won’t be surprised if he already has a fleet of vehicles for Mars in mind
2
0
29
3,678
Feature or bug, in terms of top platforms in all of tech history, X has to be the most impactful and least monetized app ever
Elon has made this place more central to public discourse than ever before, there is no other place where world leaders, CEOs, bankers, everyone cares to han X
Elon has made this place more central to public discourse than ever before, there is no other place where world leaders, CEOs, bankers, everyone cares to han X
2
1
15
1,271
An interesting consequence of mass blog generation for AEO/GEO is consumers are rapidly losing trust in AI assistants as a whole
In the past 3 mos, its somewhat obvious agents are pushing these extremely spammed tools and consumers as a result are just going back to google and esults (excluding dev tools)
Faster rollout of honest ads with a sponsored tab would replace much of this. It gives companies a separate window, and lab teams can focus on removing blogs that are factually not true
Also, there's very, very few users who are willing to trust an agent with a credit for any significant amount of money to be made. The way companies have positioned AI, it scares customers more than it educates them. Unless this changes, there's going to be very little volume
In the past 3 mos, its somewhat obvious agents are pushing these extremely spammed tools and consumers as a result are just going back to google and esults (excluding dev tools)
Faster rollout of honest ads with a sponsored tab would replace much of this. It gives companies a separate window, and lab teams can focus on removing blogs that are factually not true
Also, there's very, very few users who are willing to trust an agent with a credit for any significant amount of money to be made. The way companies have positioned AI, it scares customers more than it educates them. Unless this changes, there's going to be very little volume
7
1
33
2,793
We're in an era where most (if not all) AI businesses are funded by zero cost of capital (venture capital) so rationally it makes sense to not think about other business models than subscriptions
Like the internet, we're going to see non vc funded businesses spring up pretty soon, and for these cohorts of businesses, ads (not spammy static ads but an evolution of this) will be the only way to subsidize the massive cost that goes into serving consumers
This has happened in the past too, in the 90s everyone thought the future would be subscriptions or pay-per-use and then ads came along and monetized a ton of users
Even ChatGPT only has ~50 million paying subs. Even if you assume all of them will pay for the Pro plan at some point, that's only $10B a year (long shot). And they have massive penetration already. Some evolution of ads is the only way consumers can be monetized
Like the internet, we're going to see non vc funded businesses spring up pretty soon, and for these cohorts of businesses, ads (not spammy static ads but an evolution of this) will be the only way to subsidize the massive cost that goes into serving consumers
This has happened in the past too, in the 90s everyone thought the future would be subscriptions or pay-per-use and then ads came along and monetized a ton of users
Even ChatGPT only has ~50 million paying subs. Even if you assume all of them will pay for the Pro plan at some point, that's only $10B a year (long shot). And they have massive penetration already. Some evolution of ads is the only way consumers can be monetized
@mignanowe’re asking the wrong question. Not every consumer is the same, and there won’t be one business model.
Booking a flight is one thing. Helping a family manage kids, aging parents, health, work, and all the services around them is another.
Some agents will be free. Some subscription. But the bigger opportunity is owning the trusted relationship, combining AI with hyperlocal intelligence and integrated services, and capturing a share of the spending you help coordinate or deliver directly. Employers and health plans can fund access too.
And I suspect the most valuable agents will be a lot more human than people expect.
The real money won’t be in selling inference. It’ll be in the relationships, services, and spending those Human+AI services increasingly control.
4
1
55
3,816
No offense to the Zscaler mafia (OpenAI sales) but Anthropic sales is unfortunately much better
Playbook sales is great if a co is selling boomer saas and enterprise buzzwords but if you're at the forefront and selling the future, you need to invent new things all the time, s are very resistant to doing
Zscaler / Snowflake are amazing cos but there's a reason why their stock has been for the most part flat over the past ~5 years
Playbook sales is great if a co is selling boomer saas and enterprise buzzwords but if you're at the forefront and selling the future, you need to invent new things all the time, s are very resistant to doing
Zscaler / Snowflake are amazing cos but there's a reason why their stock has been for the most part flat over the past ~5 years
13
1
67
15,684
A lot of companies are optimizing to spend 80% less tokens when in fact they should be worried about not spending 80% more tokens
1
3
37
1,674
Varunram Ganesh retweeted
We started in the clouds at 9 W 57th 49 fl and our first factory in SoCal November 2025.
Since then we've expanded our footprint to multiple states, multiple programs of record, creating high paying jobs giving folks a renewed sense of purpose.
Excited for all on the horizon.
Since then we've expanded our footprint to multiple states, multiple programs of record, creating high paying jobs giving folks a renewed sense of purpose.
Excited for all on the horizon.
25
18
134
15,917
You either die a Post Training Researcher, or you live long enough to see yourself become the data seller
3
1
66
2,157
OH in the Marina: if you say Zach Frankel three times in front of a mirror every day, 10M ARR appears in your company Stripe one day
5
1
118
11,733
Over the next two years, we are going to see two things:
- A shift in $500B worth of ad spend from Google / Meta to ChatGPT, Muse, and other LLMs
- A new unlock in terms of ad surfaces, which will open up various opportunities for new display networks and ad exchanges
Ad Tech is back
- A shift in $500B worth of ad spend from Google / Meta to ChatGPT, Muse, and other LLMs
- A new unlock in terms of ad surfaces, which will open up various opportunities for new display networks and ad exchanges
Ad Tech is back
Consumer AI primarily monetizes via subscriptions. But <5% of consumers pay for ChatGPT, Gemini or Claude.
ChatGPT launched ads in February, crossed $100m runrate revenue in 6 weeks, and $1b by August.
Google + Meta combine for $500b+ ad revenue.
We are so early in AI ads.
6
4
47
2,763
