mojesko

HIS CI WENT RED AT 2AM. HE WAS ASLEEP. IT SHIPPED THREE PRs ANYWAY. Not a demo. A repo, five text files, $5 a night. - On-call engineer for CI: $180K a year - The loop: $5 a run - Time to build: one evening But here is why your version will fail. A prompt makes YOU the loop. Write, wait, fix by hand. It all halts the second you step away. A loop runs itself and wakes you only to escalate. The difference is not the model. It is whether something can fail the work without you in the room. Read that again. The model that wrote the fix will never call it bad. That is how you wake up to a green suite and a deleted test file. Done is exit code 0. Never an adjective. Tomorrow two people wake up. One opens three finished PRs. The other opens the repo he closed last night. Same tools. Same models. One built the gate. Save this before you build yours.

@KnightPredict glad to hear that bro!

@mojeskoqq Thanks, I'll go check out the interview

A 1,000,000-token window. The model actually holds 600,000. Sierra's Head of Product (agents for half the Fortune 20) spends 1:27 on what to do about it: 00:00 How an agent actually gets built and shipped 11:04 Meeting the model on its turf, not yours 32:22 What the agent sees at the moment of action: data, not a dump 40:00 Context engineering: everything it needs, nothing more 41:38 "Whenever you think the model's too dumb, the model's actually too smart" 46:13 Multi-agent systems are a trap: you just ship your org chart 57:03 Memory as a first-class primitive, not a patch 1:02:47 Why there is still no breakout memory company These 1.5 free hours replace 10 paid courses on context engineering. Context is not a bucket. The more you dump in, the dumber the agent gets. Watch today, then read the article above.

@shmidtqq 47 mins well spent, that agent management bit is chef's kiss

The PromptLayer founder just broke down in 47 minutes how to build agents and not drown in them: 00:00 - why teams always overbuild their first agent 03:16 - tools matter more than a perfect prompt 20:30 - how to get one agent to manage other agents 25:39 - demo: one agent launches a team of 10 28:03 - how agents prove their work BEFORE you ever open it These 47 minutes will replace 10 paid agent courses. Watch today, then read the article above on how to build a loop that fixes CI while you sleep.

@ridark_eth 🤝 thx

@mojeskoqq The interesting conversation with this speaker landed

@ridark_eth absolute gigachad move, the playbook leak is what gets me

GitHub's CEO realizing some guy replaced $1,200/year in subscriptions with free repos on 10 min of setup and automated a bot to farm free games 24/7 -> then leaked the whole playbook for free

Iman Gadzhi right after selling his $2,000 AI agents course to 500 people, watching Anthropic drop the exact same thing for free in 75 minutes

@hammertime_one bro that's either genius or about to be the most expensive lesson

A Polymarket account created 10 hours ago just put $184,552 on both England and Norway winning today's World Cup quarterfinal in Miami brand new wallet: polymarket.com/profile/0xecb8… > England YES - 52%: $126,041 → $243,602 > Norway YES - 24%: $58,511 → $246,363 if England wins: +$117,561 on that position, -$58,511 on Norway net profit: $59,050 if Norway wins: +$187,852 on that position, -$126,041 on England net profit: $61,811 Miami Stadium, World Cup quarterfinal England are favorites, but Norway have Haaland fighting for the Golden Boot and have scored in every single game this tournament winner plays Argentina or Switzerland in the Atlanta semifinal would you take that hedge?

@sama yeah the creative destruction narrative was way too doom-pilled

so far at least, i'm pretty sure AI has been net job-creating. this was not what i expected--although i was much less pessimistic than others, i thought by this level of capability we'd have seen some impact. it is possible this direction keeps going!

@nateherk this is exactly the kind of thing the world needs more of

I put a ton of work into this Claude Code course for normal people. It's completely free and 6 hours of content. I'll take you from complete beginner to actually building your own agents and automations. I put a lot of work into making this the path I wish I had when I started. No technical background needed. If you've been wanting to get into this but the whole thing feels overwhelming, this is a structured path you can follow start to finish. 👇 Link is in the replies

@xmyttle @prophetmarketai liquidity is the real moat, this actually fixes it

prediction markets have a hidden bottleneck: not intelligence. not conviction. liquidity. you can be early to a niche and still have nowhere to express the view. @prophetmarketai fixes the boring part: create a yes/no market, let the AI price it, and back your view instantly. long-tail alpha needs this. not available in the US. #ad

3/ Are you still prompting 1 request at a time? Or are you already writing the loops that prompt for you while you sleep?

2/ The whole workshop in 1 line: an agent is the worst judge of its own work. Generator and evaluator are TWO different agents with TWO different contexts. Otherwise the agent signs off on its own garbage.

Anthropic dropped a workshop on agents that run 12 hours and don't lose the plot: 00:00 - why your agent loses the plot after 20 minutes 07:55 - the Ralph loop: your first agent that never stops 14:20 - Agent Teams: agents that split the work 25:04 - the generator-evaluator contract: the agent that kills bad work 37:56 - build your own harness: the loop that fixes itself 75 minutes of free workshop. Sonnet 3.5 held 20 minutes. Opus 4.6 holds 12 hours. $200 and 6 hours of agent time shipped a playable game from scratch. This replaces 10 paid courses on agents. Watch today, then read the article below.

vector database CEO right after closing a $2M enterprise memory contract, watching a guy give a Claude agent perfect recall for $0 with a few markdown files

An Anthropic engineer on the one flaw every agent shares: close the session, and it forgets everything it learned. in 28 min he wires persistent memory onto a Claude agent, then uses Dreaming to consolidate old sessions into recall. not a bigger context window. compression that survives the tab closing. Watch, then save the 4-step build.

@cyrilXBT great read

ANTHROPIC'S AGENT GUIDE JUST KILLED THE CS DEGREE The skill that gets you hired in 2026 isn't 4 years of college, it's building agents. Anthropic published the whole blueprint for free, in 5 patterns. One block at the core: a model plus retrieval, tools, and memory. Everything else is assembly. The entire agent curriculum that matters, in 5 patterns: - Chain → split the task into steps, each one simpler than the last - Route → classify the input, send it down its own path - Parallelize → run subtasks at once, answers vote on the best - Orchestrate → an orchestrator splits the task and delegates to workers - Evaluate → one model writes, another critiques in a loop until it clears the bar Most people collect a degree and build nothing. This is 5 patterns, one block, and a project a stranger can open from a link. The honest catch: half your tasks don't need an agent. Start with the simplest workflow, add complexity only when it measurably helps. You + 5 patterns + a shipped project = an AI engineer. No degree. Save this before you enroll in another course.

everyone who grinds 9 to 5 watching one guy build a little machine that does his entire job while he sleeps and posts it online for free