In a year from now, everything will be different

Oooooo Listen to the Music get up wake up gm bullymeow.bandcamp.com/album…

More human posts. Less fake confidence. Worth thinking about

Good morning and God bless

while 500 still feels distant -- enjoying this Mac set. 1 bird observed across all 5 Sanctuaries.

成为会员，购买一只猿猴

I am in the @thearcitects genesis. seat #1604. inscribed once on ethereum. apply with your evm address ↓ ref ARC-1604B739 arcitects.fun x.com/thearcitects/status/20…

GM real ones Simple does not mean lazy Worth trying today

RUMORS: Some entities may have been developing biological weapons using Claude Fable 5. According to these claims, this is why the U.S. government requested Anthropic to shut down that model. Nobody is prepared for what's coming in the next few months. I expect total chaos across the world: economically, politically, and technologically. The signs are already there. Good luck to everyone. Stay safe!

I spent the last few weeks doing something our team avoided for years. We have legacy projects with so much technical debt that nobody wanted to touch them. Opening those repos felt like archaeology. So we ran an experiment with Fable 5: let an AI agent loose on the worst codebase and see what happens. Short version: I fell in love. The agent navigated code we were scared of. Refactored things we kept postponing for quarters. It actually worked. Then I looked at the token usage. Tokens went fast. Much faster than I expected. The more useful the agent became, the more context it consumed. Every re-read file, every re-sent context window, every retry loop was billed. Turns out we are not the only ones learning this. Microsoft ran an internal Claude Code program and per-engineer costs came out around 500 to 2,000 dollars per month. The bills consumed the team's annual AI budget within months, per Bloomberg. They are now steering their own developers back to in-house tools. The wider numbers tell the same story: → average enterprise AI spend went from 1.2M in 2024 to about 7M in 2026 → blended token prices dropped 67% in a year, from 18.40 to 6.07 per million → Goldman expects agentic workflows to drive a 24x increase in token consumption by 2030 Price per token is falling. Total spend is exploding. Both are true at the same time. In some countries, running a serious agent setup already costs about as much as a junior or mid level developer. Not replacing anyone. Just sitting there as a second salary line nobody budgeted for. I used to think cheaper tokens would eventually solve this. Now I think consumption grows faster than prices fall, and the real skill of the next few years is not prompting. It is knowing when the agent is worth keeping the meter running.

two ways i want the direction of my life to go in the long run ; > build a tech that alleviates a problem of my african people, a revolutionary app or product built to fix a major issue in my country > work with a company and expand their vertical into the african community, becoming the face of their company for years to come both paths lead somewhere meaningful and i’m gonna make sure i get there, could take years but the destination is locked in, just documenting the journey from here > if it’s tech, it has to be something that actually solves a real problem people deal with every day, not just another app for the sake of it. > if it’s working with a company, i want to be the person that brings them into the continent the right way, builds trust with the community and grows with them long term either route, the goal is the same: leave something behind that actually changed things for my people

Bitcoin starting to separate from the rest of the market again is becoming hard to ignore. While BTC continues holding above its 200-week moving average, assets like ETH, SOL and XRP still can’t fully reclaim the same level. Usually when that starts happening, it says something about where capital feels safest once confidence across the broader market weakens. Derivatives positioning also still looks relatively cautious. Long liquidations continue building, leverage appetite remains muted and traders are still paying more for downside protection than upside exposure across major expiries. The market still feels more selective than fully risk-on. Informational purposes only, not financial or investment advice.

Tag the friends who helped you through the art world. I’ll start first: @jedi6900 @suparno67

It feels like GPT-5.5 Extra High is enough for what most of us are doing, if the Pro plan keeps the same rate. I'm not sure why or for what purpose you need to seek the next models.

before there were puppets there was @lasogette

GM ☕️ Small account note good CT habits makes your content feel less scheduled. Curious how you see it

gengar ex swirled sketched raw pokemon grail here 🫠

The timeline needs more posting rhythm and less noThat is the whole point

ij8.ai introduces interactive, AI-native tutorials for creative coding. Guided by 'The Middle Way', the tutorials balance foundational computational literacy with AI as a collaborative creative partner. Each tutorial is structured as a code experiment, supported by companion theory and core principles that encourage play, exploration, and learning through making. Classroom pilots are now lining up. Please get in touch if you are interested in bringing ij8.ai into your classroom or integrating the platform into your creative practice. @ij8_labs

Good morning beautiful kings, queens, NPCs and bots! Gf coco pops and soy milky, instant coffee, kiwi fruit and banana. THANK YOU FOR YOUR ATTENTION TO THIS MATTER! #chadlife #dowhatyouwant #bewhoyouwantobe