208 Photos and videos
Nevv🗿 retweeted
Really excited to open source a new project: Omnigent, a meta-harness for AI agents.
It lets you build multi-agent coding and custom agents, sitting above Claude Code, Codex, Pi, and agent SDKs to let you compose them. It also adds live collaboration and rich control policies.
66
160
872
122,452
Nevv🗿 retweeted
Jun 12
Always check your DMs
1,955
1,242
25,598
2,139,803
Holy Moly. This dude created a plugin that lets you get paid for watching ads while you use Claude Code or Codex
Jun 11
Get paid to wait
The Claude Code spinner might be the most watched line on Earth.
So I turned it into an ad marketplace.
Advertisers bid on it. You keep 50% of the money.
Install the extension → get cash from ads.
Introducing Kickbacks
2
206
Nevv🗿 retweeted
Jun 10
It seems like most startups in San Francisco are selling products to each other
When I ask founders who their target audience is, 90% of the time it's "engineering and product teams, AI-native startups"
Feels like the same small group of target audience is being bombarded with a million products, whereas very few people are building for the 99% of the world
110
26
380
43,761
Nevv🗿 retweeted
Jun 8
The belief at the time was that model costs were halving every 6 months meaning tokens would get cheap and so application layer companies would need to find a way to charge for the value of tokens by selling the work / services.
What actually happened is AI got much more expensive than people realized at the time. The shift from chat to agents led to an explosion in cost. One user could trigger hundreds of agents and each of those agents could trigger more agents. Agents started running longer and more autonomously. On top of that frontier models like Mythos are getting more expensive not less.
If you look at what is happening in engineering, the coding companies are doing incredibly well selling tokens because engineers are consuming so many rather than needing to sell the value of the work. In fact, there is starting to be a massive demand for enterprise infrastructure to help large organizations track, manage and optimize their agents / tokens.
Now the problem for application layer companies is how do you take that large token cost and convert it into something useful for your customers. A rough analogy is every company is about to get the ability to hire infinite employees. The main challenge is going to be figuring out how to manage those employees and make your business model work the same way it did with human employees.
You can think about the previous generation of enterprise SaaS as building tools so that organizations could manage a large number of humans and make them productive. All of this is going to get rebuilt to support hybrid human / agent organizations. Some will be built by model / cloud, some by existing enterprise SaaS and some by new cos.
For Harvey this means we don’t have to become a services company. The infrastructure for every law firm to deploy, train and manage a large number of agents is going to be so complex that model / cloud providers and law firms likely won’t build all of it. And AI is going to be expensive enough that we can capture something that looks like labor spend which is much closer to services without actually having to sell services.
Jun 2
Thanks for the response, Gabe. I think people are questioning this due to a few factors. To quote Sequoia, your lead investor:
"The next $1T company will be a software company masquerading as a services firm.
Every founder building an AI tool is asking the same question: what happens when the next version of Claude makes my product a feature? They’re right to worry. If you sell the tool, you’re in a race against the model. But if you sell the work, every improvement in the model makes your service faster, cheaper, and harder to compete with. A company might spend $10K a year for QuickBooks and $120K on an accountant to close the books. The next legendary company will just close the books."
sequoiacap.com/article/servi…
25
8
279
76,326
Nevv🗿 retweeted
Jun 8
There will be an extreme irony if these models really are bound by human generated training data. RL doesn't generalize and is only useful in a handful of areas. And we all loose our skills to something that'll forever be a B player.
105
31
902
121,713
Nevv🗿 retweeted
Jun 6
This is an insane paper and I love it
arxiv.org/abs/2605.31514
158
1,304
11,217
619,426
bio/acc
Jun 5
holy shit. it happened
they just edited human embryo DNA.
this is the beginning of the end of genetic diseases.
3
144
This was literally a black mirror episode
Introducing pump fun GO: Pay ANYONE to do ANYTHING
Create & complete bounties for ANY task and leverage the power of humans & money across the globe
The world is at your fingertips. It’s time to GO 👇
2
89
This tweet is pure gold for anyone trying to grow on X
got a lot of DMs about how I hit $57k last month
short answer: 4 streams running in parallel, not one big hustle
here's how each piece actually worked
[ the 5 moves that made this work ]
1. content-engine instead of content creation
I spent 6 hours a day writing X posts manually for months. built a Python pipeline that scrapes AI/tech sources, ranks topics by virality, drafts in my voice with a verifier loop. posting volume went 3x with the same hours, brand-deal inbound went 5x. that compound feeds every other stream
2. KOL deal tracker, not a notion list
my manager and I were losing 5-figure deals every month to conversations that slipped between DMs and emails. built a real pipeline in xlsx with status pills, stage gates, and Ronin-approval flags. fixed the leak. that one fix holds most of the $40k distribution line
3. agentic workflow contracts on retainer, not hourly
priced builds hourly like a freelancer until I realized I was capping my own income. moved to monthly retainer with a clear deliverable per month. clients pay more, scope is locked, I run 2-3 in parallel without burning out. that's the $7k line and it scales
4. X monetization is downstream of bookmarks, not impressions
stopped optimizing for views, started optimizing for saves. every long-form post now has a screenshottable line, a real takeaway, a CTA that doesn't smell like marketing. went from $200/mo to $3.4k/mo in 4 months
5. paid campaigns gated by alignment, not by budget
strict filter: I only run a paid campaign if I'd post the same content for free if the product was the right fit. half the volume, twice the rate per post. that's $6.8k of campaigns I'd actually defend in DMs
[ the actual tool stack ]
- Claude Code Codex Gemini CodeRabbit for the builds
- content-engine (Python SQLite Flask Chrome extension) for X content
- KOL tracker (xlsx with status pills and brand roster)
- AGENTS.md and /retro/ folder per project for institutional memory
- Cal.com for the discovery-call funnel
- Stripe for retainer billing
[ what's different from how most operators run it ]
most people scale one revenue stream until it caps then start over with a new hustle
I built 4 streams that share an underlying system so each one compounds the others. distribution work funds the build work
build work proves the system which sells more distribution work. X content is the lead-gen layer for both. the loops feed each other
the move that compounds in 2026: stop running one thing harder, start running 3-4 leveraged streams off the same underlying system
2
111
Let's see if real usage is comparable to benchmarks. I was previously a bit disappointed with MiniMax M2.5 coding capabilities
Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities
- Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas
- MiniMax Sparse Attention scales context to 1M
- Natively Multimodal from Step Zero
API: platform.minimax.io
Token Plan: platform.minimax.io/subscrib…
🚀New! MiniMax Code: code.minimax.io
Weights & Tech Report in ~10 Days
108
Crossbar took the opposite approach.
The hardware description for their Daric chip is on GitHub.
Audit it. Fork it. Build with it. Same silicon ships in their wallet.
github.com/crossbar-inc/dari…
1
117
Publishing the design only matters if you can verify YOUR specific chip is the same.
Using IRIS (arXiv:2303.07406) short-wave infrared to see the silicon you can check chip against the public layout and verify what's inside of your @crossbar_inc wallet without ever opening it.
65
Nevv🗿 retweeted
May 23
God how soft we have become.
There are multi millionaire traders in the city making millions a year on a diet of beer and coccaine
43
57
8,510
343,063
Genie team from Google is cooking
May 19
Woot! You can now simulate real world places by grounding Genie 3 experiences with Street View imagery.
Google sitting on the mother lode of real world data, and is starting to put it to work!
Let's dive into some prompts & locations I tested...
150