Alex (@Alex_m) profile picture

Occasionally ship useful ideas

Joined April 2009
Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)Photo by Alex (@Alex_m)
Alex retweeted
- Delayed release of Astra 6.1 - 20x plan is now 10x plan - 5x plan is now 2.5x - You will pay $500 now to get the same limits as you did on the $200 plan - GPT 6 Sol is a disappointment OpenAI has completely lost it. Not even excited for DevDay anymore.
178
205
34
6,410
315,765
Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead. You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier intelligence. You’ve also claimed the lead is widening because of recursive self-improvement. I don’t see what you see in the lab. If the unreleased models are scary enough that you think you should slow down, I support your decision to be responsible. But stop pretending you need anyone else’s permission. Stop pretending antitrust law has to be suspended so you can form a cartel. Stop pretending you need a regulatory approval process that supersedes product liability. Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff. Stop pretending you need those same evaluators to police competitors who aren’t even at the frontier. Most of all, stop pretending the motivation to slow down is purely altruistic. You face massive product-liability exposure if your products enable a truly damaging cyberattack. The market already punishes models that behave in unpredictable or unauthorized ways. After the Hugging Face episode, it is simply good business for OpenAI and Anthropic to trade some raw power for reliability and predictability. Call it alignment if you want. It is also just giving customers what they want. Pacing the frontier would also create breathing room for a more intelligent conversation about regulation than Bernie Sanders’ “shut it all down.” China is very unlikely to join a global agreement, as you know, and that has to be taken into account as well. So go ahead and pace the frontier. You are the ones setting it. The easiest way not to build superintelligence is for you to agree not to build it. Demanding your preferred regulatory framework as the price of that will look like blackmail of the public and the political system. So just do it. If you do, you’ll buy goodwill for the next conversation. If you don’t, we’ll know this was just another bid for regulatory capture — or an election-season psyop.
3,373
12,217
2,709
71,685
9,649,962
Alex retweeted
the openai hugging face incident gets cited a lot but what about the detail where they had to use GLM 5.2 to analyze the attack because the "safe" proprietary models refused
162
765
66
10,596
361,068
GPT-404
35
57
5
1,080
31,494
Alex retweeted
Replying to @OpenAI
Delays after delays. Just create a cyber version and offer it to B2B/B2G and stop shooting yourself in the foot with this fear mongering mythos style propaganda bullshit that cause you more harm than good.
1
2
72
9,229
Alex retweeted
Anthropic made a generational mistake. Framing AI as a technology that’s going to take everyone’s jobs, create cyber weapons, enable autonomous warfare, etc. instead of curing disease and helping humans do more is one of the most unforced errors I’ve ever seen. How difficult would it have been to frame AI as technology so powerful it’s going to discover cures for disease, allow humans to be more productive, expand what small teams can build, accelerate GDP, and give people leverage they’ve never had before? Such an unforced error.
131
59
12
866
63,413
Alex retweeted
Replying to @AnthropicAI
@AnthropicAI At this point, the only real solution is to open-source Claude Code. The source has already leaked, so there’s nothing meaningful left to hide. Keeping it closed now only shields the team from scrutiny while users keep dealing with the same failures over and over. If Claude Code is going to keep breaking in public, its development needs public accountability too
THIS GUY LOST $200 IN ONE DAY BECAUSE THE STRING "HERMES.md" WAS IN HIS GIT COMMITS HERMES.md is a real convention used in AI agent projects. it's a system prompt specification file. not some obscure edge case he's on claude max 20x at $200 a month. yesterday claude code hit him with "you're out of extra usage" out of nowhere his dashboard showed 13% weekly usage. 0% current session. 86% of his plan was sitting there untouched but $200.98 in extra usage already burned through what should have been covered by his subscription he tried logout & login, different models, fresh installs and nothing worked anthropic support sent the ai bot (four rounds of the same scripted response). eventually they just gave up on him so he started binary searching repos and commits manually on his own time until he found the trigger the string "HERMES.md" in a recent git commit message uppercase, with the .md extension, anywhere in your commit history that's it claude code includes recent commits in its system prompt and something server side flags HERMES.md and quietly routes you off your max plan onto API rate billing > AGENTS.md? fine > README.md? fine > HERMES without .md? fine > lowercase hermes.md? fine > uppercase HERMES.md? you're getting charged API rates he reported it. anthropic support acknowledged the bug three times, called it an "authentication routing issue", thanked him for finding it then refused to refund the $200 so the man pays $200 a month for max, lost another $200 to a billing bug they confirmed, did anthropic's QA work for free on his weekend, and got a "thank you for your patience" in return check your commit history before claude code quietly drains your account too
1
7
3,288
Alex retweeted
Replying to @CodeByNZ
It just misinterpreted the request.
1
10
344
Alex retweeted
What in the actual fuck?
Would-be UFO whistleblower died of accidental drug overdose after agreeing to testify to Congress trib.al/ph0Hz9L
1
2
4,200
Alex retweeted
Replying to @mattshumer_
Anthropic's internal team was running an entirely different build from the one users got. They weren't using the same broken version everyone else was stuck with, so they missed the disaster for months. That's the explanation.
2
2
26
2,199
Alex retweeted
Let me get this straight: the people building Claude Code weren't even using the same version they charge us $200/month for. That's why it took a month of paying customers screaming into the void before anyone noticed it was broken. Cool beta test, thanks for letting us fund it.
Over the past month, some of you reported Claude Code's quality had slipped. We investigated, and published a post-mortem on the three issues we found. All are fixed in v2.1.116+ and we’ve reset usage limits for all subscribers.
1
1
7
2,595
Alex retweeted
David Wilcock spent two decades repeating one promise: "I will never commit suicide” On April 18, his co-author Wynn Free died. On April 20, a frantic 911 call reported an "unknown problem" at Wilcock's Colorado home. Police found him dead from a self-inflicted wound. Age 53. He'd livestreamed 48 hours earlier. Planning new content. Telling viewers "every day is a gift" because, "frankly, people are disappearing." Not a goodbye. A warning. Look at the roster anyway. Amy Eskridge, 34, anti-gravity researcher. Told friends she feared for her safety. Dead by gunshot in Huntsville. Monica Reza, 60, NASA JPL materials director. Vanished on a hike last June. Retired Maj. Gen. William McCasland, ran classified aerospace programs. Disappeared from his Albuquerque home in February. Left his phone. Left his glasses. Gone. Los Alamos. Air Force Research Lab. MIT. Caltech. At least 10 or 11 since 2022. This isn't heart disease. This isn't car accidents. This is scientists leaving everyday items on nightstands and vanishing into forests. Researchers who spoke up, then shut up permanently. And the FBI just now wants a multi-agency review? After four years? Wilcock wasn't a passive blogger. He sat on the board of Stavatti Aerospace. Exotic propulsion. Antigravity. The exact breakthroughs that end oil wars and make billion-dollar weapons programs obsolete. Forget aliens. Forget Ascension. Just look at the dates. His biographer died. Two days later, the guy who preempted the suicide narrative for 20 years was dead by his own hand. If this were fiction, you'd call it lazy. Too on-the-nose. But it's April 2026. And Wilcock's own last line keeps echoing: "It is just such a wild time to be alive." Tell me about it.
🇺🇸 UFO researcher David Wilcock committed suicide minutes before cops showed up, just days after he publicly said he’d NEVER kill himself. He was no scientist, mind you, far from it. Still, there has been a string of weird deaths hitting scientists and engineers working on advanced tech, and people were quick to point out the similarities. Weirdest thing of it all? He was born in Rotterdam (NY) and died in Nederland (CO). The universe really went full circle with the Dutch symbolism... Source: TMZ, Boulder County Sheriff
1
1
17
5,646
Alex retweeted
Imagine spending billions training Claude Opus just to have your own harness make it look 20% dumber.
1
1
2
20,104
Alex retweeted
Replying to @AnthropicAI
5 gigawatts. $25 billion... and you'll still deliver slop like opus 4.7 that even your own engineers can't stand. But yeah, more electricity was definitely the missing ingredient
1
25
1,044
Alex retweeted
Replying to @theo
you'd think a company building the world's smartest ai could just route the traces to one of their own models to judge if a ban is justified... instead they're out here with manual reviews and if/then workflows like it's 2005
2
1
6
1,027
Alex retweeted
Replying to @bcherny
Feeling that ~35% tokenizer tax already. Hit my 5-hour Opus 4.7 limit in under an hour on the $200 plan. 🙃
7
4
3
83
9,981
Alex retweeted
Replying to @thdxr
Its insane. It one shotted my session usage limit.
8
2
2
318
8,074
Replying to @sandislonjsak
They are lying, or the agents are poorly managed slop factories. Most of it is trash hype clickbait. I’m not anti AI just anti scam. It’s doing amazing things and the scammers are everywhere - same people who were doing crypto rug pulls now doing AI scams (courses, prompt catalogs, potions, lotions, magnets and shakes)… low level scam. Instead - I recommend read the landscape, spot an opportunity, define your goals in crystal clear HD and learn to use AI in pursuit of those goals - but do so in a way that builds your vision not your dependency on AI ( use it to build the factory, not the widget).
1
1
103
8,884
Alex retweeted
Replying to @thsottiaux
i’m genuinely worried about how often i’ll hit my weekly limits on the pro plan after the 2x increase ends on april 2nd
5
1
53
2,468
Alex retweeted
Replying to @sama
Adding the exact clauses Anthropic got blacklisted for demanding just proves they were right all along. Also, claiming the NSA won't touch your models when the former head of the NSA sits on your own board is wild. 🤷‍♂️
1
15
611