Ran (然然)

Turns out “thinking something mean but accurate” only becomes a monetizable skill once you post it. Office girl in East Asia auditing the AI circus on a $20/month research budget. Still waiting for Elon to put me on payroll.

@sama @KatieMiller Just get Premium+. Your account gets decent traffic, so the extra reach and creator payouts will easily cover the $42/month. Trust me bro.

@KatieMiller no, i pay $84/year

You post on this site for free

@sama still sorta lack of taste though

still sorta breaks my brain to see our models be good at design finally

@uday_devops Why didn’t he? I thought Vercel was free.

How did Mark deploy Facebook without Vercel ..?

the anthropic overlords won’t even grant us peasants access. “scam” altman may steal the data, but at least he democratizes the loot. our robin hood.

@sama the anthropic overlords won’t even grant us peasants access. “scam” altman may steal the data, but at least he democratizes the loot. our robin hood.

"hard questions are great but only if we deem you worthy enough to not silently downgrade you, or even get access at all"

i thought this was satire, kept looking for the handle to be spelled c1audeai or something

@TedPillows $55K and suddenly two kidneys starts feeling excessive.

Spot continues to dump $BTC. This isn't looking good.

@TheMisterTurtle they are not？

you ever get the suspicion that everyone on here is just trying to take your money

@theswansjr That’s true for pretty much every appreciating asset. Can you name one that spends over 10% of its life at ATH?

Bitcoin spends 96% of the time below it's all-time high. Surprised? Check it out here. youtu.be/4bDx3qiTyes?si…

@Amank1412 “#1 on OpenRouter” mostly means “currently free.” I tried Hy3 — Gemini 3.0 tier at best.

Tencent’s Hy3 just took the #1 spot on the OpenRouter LLM leaderboard. A 295B MoE model fully open under Apache 2.0, so it’s actually commercial friendly. Built for reliable affordable agentic workflows while going head to head with trillion parameter flagship models. Open models just keep getting better.

@Nostre_damus I’ve seen this episode before

Trump says the US will charge a toll to protect the strait The jokes write themselves

@davidpattersonx AI can do the work. It just can’t hold the license, take the liability, or go to prison when everything blows up.

By 2030, demand for all information work will be saturated. You will have as much as you want - just a prompt away. Law, medical consultation, therapy, accounting, financial services - as much as you want. Video games and movies - continuously generated in real time however you want them. All information work combined will shrink to less than 1% of the economy. Cost will be insignificant. Have as much as you want.

@ProgressBar202_ I swear every time I see your posts another 1% has disappeared. Is this year only 100 days long?

2026 is 53% complete.

@jukan05 I’m too much of a peasant for the Anthropic overlords to grant me access to Fable.

I used to think people who preferred Codex were just Dario haters, but after actually trying it, Codex is way better than CC. Opus is just too dumb.

@dev_maims Hit the limit and it’s still a senior developer. Just a 90-year-old one with Alzheimer’s who can’t remember a single function name.

Claude will make you feel like a senior developer… until you hit your usage limit. 😂

@sflorimm You can even have internet access with $0 budget

$0 budget. how do you get 1,000 paying users?

@jimcramer Make Hynix Great Again!

Walking SK Hynix back up, the American version

@teortaxesTex and the RAM price too

I need memory chip stocks to crash so I can fucking buy any

@ash_twtz That lil’ blue screen they show you every time Windows crashes is pretty reliable too.

The only Microsoft feature that doesn't break

@0x45o When you have a real job, every toilet break is paid time stolen from the capitalist class.

if you make $10k/month, every 10 minute toilet break costs you about $2.31 this is a serious problem that needs to be solved

@GergelyOrosz means to type “/goal” before the actual prompt？

What is "loop engineering" to you, anyway? Been looking into this, and also experimenting myself... and I'm not (yet?) buying that it's a new paradigm. But I might just be missing something. Can you share "loops" you regularly use?