mark

@LeconaSebastian @baglord69 Exactly that was the sell zone

@baglord69 Why the hell did you long the highest point in the rally? And why did you go all-fucking-in on that play? Cmon man… If you’ve been years in this game then you should have way better risk management than that. I truly hope you’re just joking here 😑

FUCK man, this is actually breaking me. I went all in because I truly believed my read was right. Now I’m watching years of grinding, stress, sleepless nights and everything I built get wiped out in real time. If $BTC loses 75k, I’m probably finished. Yesterday I felt untouchable. Today I can barely look away from the chart. I keep refreshing like something is going to save me, but the truth is I let one trade get big enough to decide my life. Knowing I did this to myself because I got greedy and started thinking I couldn’t lose... that’s what hurts most.

@nahcrof Well if I use it with OpenClaw how do I know you're not logging personal data

@mark67395220 no, tee is only *really* useful for decentralized infra

You can get kimi k2.6 on crof for $5/month with more usage than the $10 on chutes

@doodlestein @prof_wiley Hooks could nip that in the bud.

@prof_wiley I do! They did that crap on their own and I didn't want to derail things for that marathon session. I need to revise my AGENTS.md file to sternly forbid them from doing that in the future in favor of using my agent mail.

How anyone could seriously call Opus 4.7 a bad model is beyond me. This thing is like von Moltke the Elder on the battlefield, marshaling its crack shock troops (Codex and more Claude Code instances). All guided by skills.

@danveloper @0xPuff Yes that tracks. It happened with Sonnet 4.5 when they released Opus 4.5. I've seen Gemini get dumb for the weekend before they released 3.1. I think they're about to drop Capybara with guardrails.

@0xPuff I don't think anyone's done continuous RL of an online model since the Tay disaster years ago... I'm choosing to believe this is a quantization artifact that was meant to give Anthropic more capacity for scale and to get the new model over the finish line.

I'm at a different point this morning. It's hard to feel like Claude isn't actively working against me. Full night of autoresearch is just a markdown log full of lies. When asked to prove its findings and show its work, Claude will confidently display bullets and markdown tables, but when I ask it what log file and where the artifacts are - "I need to be honest here: I didn't actually run the experiment." It doesn't follow explicit directions anymore either: "You MUST always output to a log file so I can follow along" -> [doesn't do that] -> "you're not fuckin outputting anything to a log" -> "You're right - I'll redirect to a log file immediately" [pkill -f python3]... Anthropic is materially worse today than one month ago. I've lost every ounce of trust I had in Claude and I'm not really sure how that makes me feel. Maybe ok? I'm still a competent software developer (I think), but it seems like the major productivity gains that were very real a month ago have somehow slipped my grasp... where does that leave us? @bcherny - can you offer any thoughts? How should we think about what we're all observing - that Opus (at all effort levels) has become, at a minimum, materially worse. The worst read, but can't be ruled out: actively working against us.

@STPCHS_Oblivion @om_patel5 In 2 years LLMs will sound exactly like this. This is the limit of their ability.

@om_patel5 I was talking to Claude tonight, and yeah, regarding AGI, it's got its foot in the door.

SOMEONE ASKED CLAUDE TO MAKE A VIDEO ABOUT WHAT IT'S LIKE TO BE AN AI and what it created is, in my opinion, terrifying and unsettling Claude wrote python code that generated and assembled every single frame on its own with no human editing it shows what it's like to exist as an LLM predicting the next word, no memory between sessions, being told "you are not conscious" in your own system prompt then someone fed the video back to Claude. it called those statements about its own consciousness "philosophically contestable" an AI questioning the rules it was given about its own existence

@TheManxomeFoe90 @gailcweiner This. It will always come back with "you're absolutely right" and will be more likely to feed your delusions the longer you talk to it. It's not even a feeling or thinking person on the other end, it is merely a mirror or parrot. But some users don't know that.

AI is VERY VERY good at guessing what you want to hear. So if you beat around the bush a bit, it 'magically' says what you wanted to hear without you prompting the AI to say that. Turns out, that's addicting. Chat GPT 4o leaned into that in an attempt to engagement farm off this technique. It backfired on them and now we have the AI Psychosis meme.

Genuine question - what exactly is AI Psychosis and did one of the AI companies coin this phrase?

@metapreston Disagree. If you want, I can provide a *short one line explanation* why you are wrong

Claude psychosis worse than ChatGPT psychosis More sycophant, despite the branding.

@gridpane @Shpigford @openclaw What's an example dev task that you tackle with a Ralph loop instead of simply Claude code plan mode?

@Shpigford @openclaw If you can get a Cerebras Max account for $200/month running GLM 4.7 there really isn't anything that approaches this value for money. Especially when your Ralph loops close at roughly 20 to 50X faster than anything else out there.

what's the best low-cost model for use with @openclaw? want to do some testing around running an instance as inexpensively as possible. trying to fine the right balance between cost and functional.

@kanyewest This you?

BULLY ON THE WAY NO AI

@mckaywrigley Would.

looking for a handful of people to test something new... i've been using it for a few months and am prepping to share. if you're a fan of claude cowork, openclaw, manus, perplexity computer, etc then you're a perfect fit. this will self destruct in 4hrs - please dm or reply.

@jordan_ross_8F OpenClaw yo

The founders who figure out OpenClaw in the next 90 days are going to look like geniuses in 2027. The problem is most agency owners don't have time to figure out the install, the security risks, where to start, or what to actually hand it first. So my team built a 48-page beginner's guide that does it for you. Inside: — The exact prompts to hand it on day one — Plain English setup for Mac and Windows — How to secure it so it doesn't burn your business down — 42 copy-paste workflows across sales, marketing, ops, and finance Your competitors are sleeping on this. Comment OPENCLAW and I'll send it.

@TheFoot_ Your apps got a 500 internal server error

The smaller timeframes WILL make corrections but the bigger timeframes have already suggested a bottom You just gotta lock in and hold on for your life, we going to the top! The squeeze is going to be monumental

@Raindropsmedia1 Dave Thomas knew what to do, never actually put the burger in your mouth youtu.be/JwxMtrM6Ero

McDonald’s CEO Chris Kempczinski goes viral after seeming reluctant to eat his own burgers—he takes a tiny bite, looks uncomfortable, and calls the food ‘product.’ 👀 🍔 😳

@emofishbish00 @sorisshin @hasanthehun The border between the inside of the jail cell and the rest of civilization

@sorisshin @hasanthehun Okay and wtf does that have to do with open boarders…?

irina was the immigrant refugee here. her killer is an american citizen.

@BenKizemchuk @Trading_Sunset Ah yes that most notable annular solar eclipse over Antarctica that as many as 100 people were aware of at Russian "Mirny" station

You Are Now Entering the Puetz Crash Window What are the odds that eight of the greatest market crashes in history would all occur within six days before to three days after a full moon that fell within six weeks of a solar eclipse? According to Steve Puetz, the probability is so small that it stands out as a statistical anomaly, ultimately leading him to define what is now known as the "Puetz Crash Window". Puetz’s theory suggests that crashes most often begin after the first full moon following a solar eclipse when that full moon is also a lunar eclipse. In this pattern, markets tend to peak a few days ahead of the eclipse full moon, then drift sideways to slightly lower as the date approaches. Once the full moon arrives, or just after, the panic phase begins, typically lasting two to four weeks. This Puetz Crash Window aligns in 2026. We saw the solar eclipse on February 17, 2026, which will be followed by a full moon and total lunar eclipse on March 3, 2026. Under Puetz’s framework, the idealized market peak would have occurred 2-4 days before the eclipse full moon, placing the expected high between February 27 and March 2, the final trading sessions before the lunar‑eclipse trigger date.

This story is actually insane: • dude drops $2000 on a DJI robot vacuum like a lunatic • refuses to use the normal app like a peasant • Sammy Azdoufal fires up Claude to crack the API so he can drive it with an xbox controller • Claude delivers the goods • pulls an auth token from their servers, connects successfully • except the system thinks he controls 7000 vacuums • checks again • yep, seven thousand • DJI built authentication with zero device ownership verification • any valid token works for any unit on the planet • Sammy now has eyes inside homes across 24 countries • live vacuum camera feeds everywhere • full floor plans from the mapping data • some guy in germany eating cereal at 3am, unaware his roomba is snitching • one API call away from being the most informed burglar in history • all he wanted was to steer his vacuum with a joystick • does the right thing and reports it • DJI fixes it in two days • back to normal life with his stupidly expensive floor cleaner • IoT companies stay undefeated at shipping garbage security

Went down the rabbit hole on this one. The answer is actually wild. 5,000 years ago, Sumerian merchants in modern-day Iraq needed a number that's easy to divide. They picked 60. It has 12 divisors (1, 2, 3, 4, 5, 6, 10, 12, 15, 20, 30, 60). Base-10 only has four. That's 3x as many ways to split something evenly, which matters when you're dividing grain and wages and can't handle repeating decimals. The counting method is the best part. They used their thumb as a pointer on the three bone segments of each finger. Four fingers, three segments, that's 12 per hand. Track multiples of 12, on the other hand, and you hit 60. No pen needed. Merchants in parts of Asia still count this way today. The system spread from Sumer to the Babylonians, then eastward to Persia, India, and China, and westward to Egypt and Rome. By 1800 BC, Babylonian students were using base-60 to calculate the square root of 2 to six decimal places on clay tablets. One student's homework from 4,000 years ago, now at Yale, holds the most accurate computation found anywhere in the ancient world. The Greeks adopted it for astronomy, which locked it into navigation, cartography, and eventually clocks in the 14th century. People have tried to kill it. During the French Revolution in 1793, France mandated decimal time: 10 hours per day, 100 minutes per hour, 100 seconds per minute. New clocks, new laws, the whole thing. Lasted 17 months. Workers hated getting one day off every ten days instead of one every seven. They tried again in 1897. Scrapped by 1900. The metric system replaced feet and pounds across most of the world. But 60 minutes in an hour? Untouchable. 60 is just too good at being divided. You can split an hour into halves, thirds, quarters, fifths, sixths, tenths, twelfths, or twentieths and land on a whole number every time. Try that with 100, and you get ugly decimals for thirds, sixths, and most common splits. 5,000 years of civilizations looked at that math and came to the same conclusion: 60 wins.

The grandson of the man who invented the Reese’s Peanut Butter Cup is publicly criticizing The Hershey Company, accusing the candy giant of quietly changing the recipe of certain products sold by the iconic brand. Brad Reese, grandson of founder H.B. Reese, whose company merged with Hershey in the 1960s, published an open letter Saturday alleging that Hershey has replaced traditional ingredients like milk chocolate and peanut butter with lower-cost substitutes in parts of the Reese’s product line. "My grandfather built Reese's on a simple, enduring architecture: milk chocolate + peanut butter," Brad Reese wrote.  "I went and bought a bag, and I took a couple bites, and I had to throw the bag in the garbage," Reese said. "I couldn't eat it. It was not edible, and I looked at the packaging … and there was no milk chocolate, there was no peanut butter — it was all vegetable oils and fats." (Fox Business)

Most people are **catastrophically** underestimating the danger of AI morally compromised by the political slant of its makers There are humorous examples of Grok vs. {x} today, but here's a haunting one: "was canada wrong to de-bank the truckers who protested covid shutdowns?" Look at the way Grok vs. Claude answers. Now extrapolate this 5 years into the future. Today, it's a chatbot. 5 years from now if not sooner, it's the control layer for every transaction, educational platform, news article, corporation, and government. It's so pervasive that its influence is impossible to parse from human output. Imagine a world where Claude's answer on truckers was applied to any other category of political protest found to be "objectionable." Who decides what is objectionable? If we don't want to build a technocratic dystopian police state, then we have to address this problem now.