Ryan

ok then cure cancer you fucking dorks

@cathrynlavery @bcherny I notice it does it anytime it's outside of the current project directory. Maybe they just apply to the directory it's spawned in?

why does claude --dangerously-skip-permissions not work anymore? getting asked for permissions nonstop 🫠 @bcherny

@TheAIPulseM @quionie This here is a perfect example 😂

The biggest giveaway isn't the grammar; it's the Lack of Risk. ⚖️🚫 AI writing is 'safely beige.' It avoids strong, polarizing opinions, weird metaphors, or that specific 'human' cadence of starting a sentence with a punch and ending with a whisper. If every paragraph has the same 'weight' and perfectly transitions like a corporate brochure, you’re reading a silicon brain. Humans are messy; AI is Symmetric.

what’s the easiest way to spot AI writing? besides the em dash lol

@nypost The irony here.. > Fraud the taxpayer > Ruin your image > Fraud the taxpayer to fix image

Gavin Newsom to spend $19M in taxpayer funds on New York PR firm to polish California's image trib.al/t39xDzL

@boneGPT My guy, just read the timestamps in the log file 😂

Sam Altman reacted to this video after my post lmao We are the media now

@om_patel5 This is low key genius

I taught Claude to talk like a caveman to use 75% less tokens. normal claude: ~180 tokens for a web search task caveman claude: ~45 tokens for the same task "I executed the web search tool" = 8 tokens caveman version: "Tool work" = 2 tokens every single grunt swap saves 6-10 tokens. across a FULL task that's 50-100 tokens saved why does it work? caveman claude doesn't explain itself. it does its task first. gives the result. then stops. no "I'd be happy to help you with that." no "Let me search the web for you" no more unnecessary filler words "result. done. me stop." 50-75% burn reduction with usage limits getting tighter every week this might be the most practical hack out there right now

@bcherny @Hesamation It's smart by anthropic to set the default effort lower. Most users won't switch to max, or don't require it. I definitely notice a difference in quality if I forget to switch, so I'm sure this is why people are saying it's nerfed.

👋 This is false. We serve exactly the same models to all users. What the person in the post might be experiencing is a lower effort level vs. what the enterprise set. Claude Code users can change this anytime by running /effort. low effort = less tokens and lower intelligence, high = more tokens and more intelligence.

Redditor claims Claude Code is nerfed for Pro/Max users vs Enterprise customers and the strategy is to use the paid plan users to generate hype on X and LinkedIn so companies would reach out to them.

@nypost Here's an idea, end the IRS and turn them into fraud investigators.

California's fraud problem is so bad the DOJ will 'never have enough' prosecutors to fight it trib.al/wd230RT

@goodforbill @USAttyEssayli @VP That's also my assumption

@imRyanDietz @USAttyEssayli @VP My guess is that the pols are recipients of large campaign donations from the fraudsters.

Today, we charged 15 defendants with healthcare fraud totaling more than $50 million. This was the first major arrest operation in coordination with @VP’s Fraud Taskforce. The fraud train is coming to an end in California and across the country.

@davidonchainx It will just breed more competition, and the market will equalize.

What if Anthropic's plan to profitability is: > Launch Claude Code > Make it easy for people who can't code to build apps > Everyone loves it > Ship new features every week > Every new feature burns through more credits > Users hit their credit limits faster > Soon their entire business runs on Claude > Then raise prices a lot > Everyone forced to pay otherwise their business dies > They can't leave because they don't know how to code > Anthropic revenue increases by 10x overnight They're not selling an AI tool They're selling a dependency Genius

Something I don't see talked about when vibe coding.. I used to be like "yeah I'm not building that, it's too much work".. Now that I don't have that problem, I can let the creative juices flow, without fear of burnout. The only limit is imagination.

@Jason How about no taxes until they get this cleaned up, and people at the top go to jail.

We don’t care which side eliminates government fraud — we only care that it’s eliminated. No new taxes until you audit everything and lock up the crooks!

Logo designers are safe from AI for a while. Their most is massive.

I see a lot of complaints about Claude code usage limitations.. I don't get it.. I have 2-3 sessions running simultaneously, every day, about 8-10 hours a day, and have never been limited. Is it just the $20/mo plan thats limited? I do have the $200 max plan, but wouldn't think its that much different.

@elonmusk Plus it's not like any users asked for it, it's annoying AF. I left openAI for that reason, I just want a model to give it to me straight.. @grok is much better at just shooting the breeze, researching or working through issues.

Major problem

@JesseWilsonUSC @cryptotussy @StraughterG Thx gpt

@cryptotussy @imRyanDietz @StraughterG The leak that occurred yesterday, March 31, 2026, was indeed the latest and greatest version of the Claude Code source code, not an outdated or 1-year-old version. Version: v2.1.88 (the current production version at the time).

claude code signs every API request with a cch= hash computed in compiled zig. recompiled clients just send zeros nobody knows the signing algorithm yet. which means anthropic can tell which requests aren't from the legit client. currently decompiling the official binary to reverse this - would be huge for all third party clients like opencode, openclaw etc to fully bypass anthropic enforcement and actually use the tokens you're already paying for [right now using every client which is not the official version of claude code with your subscription token CAN get you banned cause anthropic can detect it]

@jaymegordo @StraughterG True, any clients that didn't update would be at risk. I think the native installer auto updates, but if they installed with npm they could be invalidated.

@imRyanDietz @StraughterG I think then they’d have to “invalidate” all previous versions

@MelkeyDev Oracle is claiming it's to invest in AI data centers, imo I think it's a fancy way of saying they've been replaced by AI..

🚨Holy shit - 30,000 Oracle employees laid off🚨