CEO @ ShineOn, over +5m orders shipped 💸 | AI-powered entrepreneurship will save the world.
Sometimes emails 👉
Joined January 2012
- Tweets3.5K
- Following667
- Followers873
- Likes1.8K
Fable 5 = sees the forest
Sol 5.6 = sees nothing but trees, beautiful magnificent trees, and will explain - in the most excruciating detail - how all the beautiful magnificent trees harvest photons from 93 million miles away
In a coffee shop.
Not a single person with Claude or Codex up.
We are in a bubble. The world still does not know.
It 100% is until open source catches up
Even then, expect chaos
The gov is alarmed that the waves are pulling out, and have yet to understand that tsunami of intelligence is coming
Prediction:
Before 2030, the smartest AI on earth will live in a SCIF, the second smartest will work for Blackrock.
… and we’ll get whatever’s left.
I’m not being doomerist… this is already happening.
The White House is prepping an executive order that will require federal review of new AI models before public release.
Let me translate that for you:
The government is about to decide which models you’re allowed to use.
This was predictable (and inevitable) from the start.
Intelligence is dangerous, and superintelligence (where all of this is heading) is a weapon.
AI companies are essentially building atomic bombs that can think.
The public doesn’t have access to hand grenades, bio-weaepons, or nuclear bombs.
The same logic will apply here.
So, AI is about to split into (at least) 3 tiers:
1️⃣ FIRST:
The government gets the top tier.
They’ll use it for military planning, bio-weapons research, cyber operations, and a whole lot more we can’t even imagine.
There will be models the public will never see or be told about.
They’ll live in a SCIF. Classified.
2️⃣ SECOND:
Enterprise will get the next tier.
They’ll use it for massive compute of proprietary data, R&D that smaller companies can’t match.
Think Blackrock, Pfizer, Lockheed.
(Hey - maybe we’ll finally be able to reverse engineer UFO tech 😉)
3️⃣ THIRD:
The rest of us get whatever’s left over…
Stratified by income level.
The rich will pay for the sharpest assistants, best forecasters for their investment portfolios, more powerful reasoning, etc.
They’ll be able to make better decisions faster.
Their kids will learn faster.
Their investments will compound faster.
The middle class will get something competent.
It’ll be useful for everyday life, but capped.
Unfortunately, the lower class will get whatever’s free.
Eventually, a countries will nationalize their own models and make them public utilities.
Free AI for every citizen…
Like a digital library card.
---
I’m still very bullish on AI… that was never the question.
I just think the next 4 years (maybe less) decide whether AI becomes a public good or a private weapon.
It’ll almost certainly become both
Wow, this is legitimately insane
It’s weird how many successful entrepreneurs (people in general?) have a level of self-belief that borders on delusional.
If 50% of this is true things about to get crazy
🚨BREAKING FRONTIER MODEL NEWS
gpt-6 set for release april 14th
altman's team has been leaking like a sieve lately, here's what openai staff are saying privately.
>pretraining completed march 17th. post-training and red-teaming already done. this thing is ready.
>benchmarks are absurd. outperforms gpt-5.4 by 40%+ on coding, reasoning, and agentic tasks.
>natively multimodal from the ground up. text, audio, images, video one architecture
>openai killed sora and redirected every GPU to this model. the billion-dollar disney deal is dead. that's how serious this is.
>product org officially renamed to "AGI Deployment." it’s agi time baby.
>brockman says AGI is 70-80% achieved. internally they think gpt-6 closes most of the remaining gap.
>2 million token context window. double what gpt-5.4 offered.
>priced at $2.50/$12 per million tokens. barely above gpt-5.4. so like mythos intelligence, but you can afford it.
>safety team moved under the CRO. altman stepped back from safety oversight entirely to focus on data centers.
>openai has been in internal "code red" since december 2025. this is their answer.
>powers the new desktop "superapp", chatgpt, codex, and atlas browser merged into one agent.
the potato is cooked.
spud is agi.
I’m A-OK with this if it means Claude Code users get usage back 👌
OpenClaw (no hate) was highly inefficient and burned 10-gajillion tokens for simple tasks. Them internet pipes are gonna be wide open tomorrow
Now up to 3,500 posts
A masterclass in what NOT to do in customer service.
Here's the play-by-play:
03/23 - Users report hitting usage limits in just a few prompts. Looks like a bug.
03/25 - Issue goes viral on X and Reddit. 1k+ posts. Zero acknowledgement from Anthropic.
03/26 -More reports pile up. Anthropic announces that they are throttling 5-hour session limits during peak hours. Still no acknowledgement.
03/30 - A Reddit user reverse engineers the Claude binaries and finds 2 bugs causing usage issues. Hours later, Anthropic finally acknowledges the reports. Says they're "actively investigating."
---
It's now 04/02... 11 days after this started...
Users are still experiencing issues. No further updates from Anthropic.
Meanwhile, people are rage-quitting Claude for OpenAI's Codex.
I friggin' love Claude, but THIS is brutal
A masterclass in what NOT to do in customer service.
Here's the play-by-play:
03/23 - Users report hitting usage limits in just a few prompts. Looks like a bug.
03/25 - Issue goes viral on X and Reddit. 1k+ posts. Zero acknowledgement from Anthropic.
03/26 -More reports pile up. Anthropic announces that they are throttling 5-hour session limits during peak hours. Still no acknowledgement.
03/30 - A Reddit user reverse engineers the Claude binaries and finds 2 bugs causing usage issues. Hours later, Anthropic finally acknowledges the reports. Says they're "actively investigating."
---
It's now 04/02... 11 days after this started...
Users are still experiencing issues. No further updates from Anthropic.
Meanwhile, people are rage-quitting Claude for OpenAI's Codex.
I friggin' love Claude, but THIS is brutal