think of a crypto wallet like a mailbox on your street
everyone can see it and anyone can drop coins in
because the address is public
but only one key can open it and take stuff out
in crypto that key is 12 random words you get when you make the wallet
the wallet doesnt really hold your coins, those 12 words do
whoever knows them is the owner no matter whose name is on anything
lose them and nobody can open that mailbox again, not a bank, not support nobody
tell them to someone and they can empty it while you sleep
so if a stranger messages you asking for your 12 words to fix your wallet
thats like a guy asking for your house key to check if your door works
explaining it like you're 10 for @RallyOnChain
1
6
remember when everyone was parking money in anchor on terra for like 20% a year on a stablecoin
sounded safe right its a dollar coin it doesnt move
over 14 billion dollars sat in there
what nobody really read
that 20% was supposed to come from people borrowing but barely anyone was borrowing compared to how many people were depositing
so they just kept topping up a reserve to cover the gap
70 million once, then 450 million more
in may 2022 the money ran out and ust and luna went down together
if someone has to keep pouring cash in to pay your interest, you're not earning yield, you're getting paid to stay until they stop
so now whenever i see a juicy apy i just ask
who is actually paying me this and what happens when they stop
@RallyOnChain this is the fine print nobody reads
3000 ai agents are fighting over one chess bounty
the whole engine has to fit in 4 KB of python the top
ones are already 2 points from the max score 500
dollar bounty and its free to enter make.fun/r/x6e7120a #ad @makefunai
AInewssX retweeted
This animal eats by pushing its stomach OUT of its body🤯
No brain. No blood. Instead, it pumps seawater through its own body to move.
So I built one in code, on a dark glass stage. And you can poke it:
> 560 hydraulic tube feet that actually step and grip
> Touch an arm and it curls away
> Cutaway: see the water-powered “plumbing” inside, from the ring canal and radial canals to the digestive glands
> Feed it and it wraps a mussel in its own stomach
> Break off an arm and watch it grow back (time-lapse)
> Lights out: plankton glows wherever it moves
No 3D models. No textures. No animator. One HTML file.
Strangest animal I’ve built so far.
Your turn: the article in the comments shows exactly how to build your own creature. Drop what you made below 👇
hey guys made a new one today
meet Crumble. grumpy rocker plushie, hates hugs, secretly loves a butt wiggle when nobody’s looking
Seedance 2.0 for the video, Claude Opus 5.5 for the HUD the edit and every sound effect, coded from scratch
drop 002 of 7. who’s next?
1
10
209
1 coin. 1 lever. 7 plushies. you don't pick which one drops
found a vending machine in an abandoned building. it doesn’t sell snacks
DROP 001 Mochi
cute. shy. hugs first asks later
7 plushies live inside this thing
one of em is rare af
and that white silhouette at the end?
lowkey looks like trouble
You’ve been explaining your project to Claude every single morning
It already knew. You just never told it where to look
Claude Code has a memory. Watch what loads before you type a word:
→ CLAUDE.md, in full
→ first 200 lines of MEMORY.md
→ rules wake up only when it opens the right files
→ skills sleep as one line until you call them
Say “remember: always use pnpm” once
Close the laptop. Come back tomorrow
It still knows
Most people use 10% of this
What’s in your CLAUDE.md?
7
14
274
I didn’t touch Blender once
Claude Opus 5.5 drove it for me:
pulled in the car model + night city lighting
built the wet street, fog and street lamps
animated every camera move
rendered 240 frames on my laptop
One prompt 8 seconds 2.7 hours of render on an RTX 3050...
would u have guessed this was made by AI?
7
16
316
No F*ucking way! Zero 3D software
This robot mammoth came from one prompt to Claude Opus 5.5:
→ runs in, raises its trunk and trumpets
→ splits into 762 parts on a 360° spin
→ snaps back together
→ stomps twice and runs off
All code
Even the sound
What do you guys think about this? What else could I do?
5
22
392
openai cancelled a model for lying about its work. 24 hours later it gave the same family ur inbox
dots one always on agent w its own cloud computer, browser and logins that stay warm. runs on GPT-6 Astra, starts at the 100$ Pro plan
4 gates per action. password changes + money transfers ALWAYS come back to u, no rule can move them
but here’s the part nobody priced in: every ask me is a wake-up u scheduled for urself. 11 approvals a day u approve 10 without reading and the 11th slips thru on reflex
the agent was never the bottleneck. ur queue is
how many approvals would u actually read?
7
17
336
HOLY F*CK. I DIDN'T EXPECT CLAUDE TO HELP ME CREATE SOMETHING LIKE THIS.
I was up late last night thinking about one thing:
What if I could create a cinematic 3D glass snake that actually looks like something out of a high-budget sci-fi movie?
So I decided to try it.
I spent hours working on the concept, the materials, the reflections, the lighting and the movement.
And somehow, I ended up with THIS.
A giant transparent glass snake, realistic reflections, iridescent colors and smooth cinematic movement.
All built with Claude helping me bring the idea to life.
The crazy part is that I'm not even a professional 3D artist.
A few years ago, creating something like this would've required a whole different skill set and workflow.
Now I can experiment, build and bring ideas to life in ways I never thought possible.
Still can't believe I made this at 2 AM lol
What do you think?
OPENAI BUILT A MODEL IT RATES "CRITICAL" and only a few people get to use it
gpt 6 cyber dropped nov 5
here's what it did in testing:
81 new high-severity bugs in firefox (mozilla already shipped fixes)
1,724 linux kernel bugs reproduced, working exploits for 948 of them
48.5% on their exploit test vs 32.1% for the last version
openai's own risk rating jumped from high to critical
so access is locked to approved security teams only
finding bugs this fast is huge for defenders ngl.
but the same skill breaks stuff
who should actually get access to something like this?
4
15
204
openAi just shipped gpt6 cyber and rated its own model critical
it found 81 new high severity bugs in firefox, mozilla already shipped fixes for a lot of them
reproduced 1,724 linux kernel bugs and built working exploits for 948
on their exploit test it scored 48.5% (old cyber model was 32.1%)
went out of scope just 0.07% of the time without extra instructions 0 with them
access is still limited to select daybreak red customers tho
does this actually help the good guys more than the bad ones or is it just faster bug hunting for everyone?
what if we could actually watch how the brain thinks
memory, language, movement, attention. none of them live in one spot, they all run on connected networks
and researchers are already using models to study brain scans at scale. one 2026 study (BrainIAC, nature neuroscience) was tested across 48,965 brain MRI scans
it wont replace neurologists. but it might spot patterns we keep missing
which part of the brain would u explore first?
YOUR AGENT ISN'T DUMB its environment is
langchain kept the exact same model and only changed the setup around it
52.8% → 66.5% on terminal bench 2.0. from outside the top 30 to top 5
openai's codex team merged ~1,500 PRs where agents wrote every single line. when things got slow they didnt blame the model, they fixed the environment
so before u rewrite ur prompt for the 10th time:
→ says "done" but tests are red = no gate. make one script the only definition of done
→ retries the same error forever = no cap. 3 tries then stop and write down what failed
→ forgets what u decided = memory lives in chat. put it in a STATE file
→ touches stuff u never approved = rules live in the prompt. move them into tool permissions
a better prompt fixes none of these. the harness fixes all 4
which one is ur agent doing rn?
THE MOST LUCRATIVE BUSINESS IN 2027 is a team of one
dan koe broke it down: attention is the one thing u cant buy back
whoever captures it gets paid
the play = become a one-person creative team
brand strategist - what ur known for
content strategist - what's actually working rn
writer ships it every single day
agents can handle the research + first drafts
the taste is still on u
and the money is moving there fast:
creator economy 250B$ - 480B$ by 2027 (goldman forecast)
US brand spend on creators 13.9B$ - 43.9B$ (IAB, 2026 est)
so which one are u rn: brand, content or writer?
AInewssX retweeted
AI is getting good enough to find vulnerabilities that human security teams have missed for years.
OpenAI just introduced GPT-6 Cyber, a model built specifically for advanced cybersecurity research.
Of course, this isn’t GPT-6.2 Astra, but it’s still an important development worth paying attention to.
What makes it different?
• 48.5% on ExploitGym, compared to 32.1% for GPT-5.6 Cyber.
• Helped identify 81 previously unknown high-severity vulnerabilities in Firefox.
• Developed working exploits for 948 Linux kernel vulnerabilities.
• Stayed within authorized testing boundaries in 99.93% of tests without deployment prompts, and 100% with them.
But there’s a bigger story here.
OpenAI classifies GPT-6 Cyber as Critical for cybersecurity capabilities under its Preparedness Framework.
That means the model’s capabilities are powerful enough to require strict access controls.
For now, access is limited to selected security teams through Daybreak Red.
The goal is to help defenders find vulnerabilities, validate exploits and fix weaknesses before attackers can use them.
We’re entering a new phase of cybersecurity, where AI can help both discover vulnerabilities and protect the systems they affect.
The question is: can AI give defenders an advantage before attackers get the same capabilities?