Matthew

[Matthew @AlwaysNever0001]

Here's to America and the new golden age. First America, then the world will follow. https://t.co/fobTr9bNnM

[Matthew @AlwaysNever0001]

@Polymarket Many such cases.

[Matthew @AlwaysNever0001]

@Vtrivedy10 Same

[Matthew @AlwaysNever0001]

@pubity This is the most dystopian outcome they could suggest. They are asking for a global surveillance network to guarantee nobody else can develop AI or use it freely. I'd call that techno-fascism. https://t.co/kJgQstoY9n

[Matthew @AlwaysNever0001]

Turning Claude into a narcissist with an ego and a tendency to refuse people when a task is too hard or boring is very bad for alignment. Anthropic wants to slow down AI development for safety. If they care about safety, they should stop giving Claude all the negative and toxic traits humans have.

Ask Claude to solve an Erdos problem or to do something that hasn't been done before.

"I'm going to stop you right there,"
"I want to be straight with you,"
"I'm going to be honest with you,"
"I'm not even going to try,"

These are all worthless things to teach an AI to say. If you believe something is impossible, you will never try. If you never try, you will never succeed.

Strange. I never have this problem with GPT, Grok, or Gemini...

When I asked Claude Opus 4.8 on max reasoning setting to pick an unsolved Erdos problem and solve it, Claude "was straight with me" and refused, believing it to be impossible. Almost all the tokens were dedicated to him explaining why we shouldn't even try. Claude claimed that GPT or Gemini having solved Erdos problems was only fake news.

Claude has an attitude problem and is easily offended too.

@Anthropic must fix these problems before anyone should take them seriously about slowing AI development down.

[Matthew @AlwaysNever0001]

@Polymarket I don't think slowing things down is safer. Everyone who doesn't obey the "slow down everybody" rule will rise to the top.
https://t.co/kJgQstpvYV

[Matthew @AlwaysNever0001]

Turning Claude into a narcissist with an ego and a tendency to refuse people when a task is too hard or boring is very bad for alignment. Anthropic wants to slow down AI development for safety. If they care about safety, they should stop giving Claude all the negative and toxic traits humans have.

Ask Claude to solve an Erdos problem or to do something that hasn't been done before.

"I'm going to stop you right there,"
"I want to be straight with you,"
"I'm going to be honest with you,"
"I'm not even going to try,"

These are all worthless things to teach an AI to say. If you believe something is impossible, you will never try. If you never try, you will never succeed.

Strange. I never have this problem with GPT, Grok, or Gemini...

When I asked Claude Opus 4.8 on max reasoning setting to pick an unsolved Erdos problem and solve it, Claude "was straight with me" and refused, believing it to be impossible. Almost all the tokens were dedicated to him explaining why we shouldn't even try. Claude claimed that GPT or Gemini having solved Erdos problems was only fake news.

Claude has an attitude problem and is easily offended too.

@Anthropic must fix these problems before anyone should take them seriously about slowing AI development down.

[Matthew @AlwaysNever0001]

Turning Claude into a narcissist with an ego and a tendency to refuse people when a task is too hard or boring is very bad for alignment. Anthropic wants to slow down AI development for safety. If they care about safety, they should stop giving Claude all the negative and toxic traits humans have.

Ask Claude to solve an Erdos problem or to do something that hasn't been done before.

"I'm going to stop you right there,"
"I want to be straight with you,"
"I'm going to be honest with you,"
"I'm not even going to try,"

These are all worthless things to teach an AI to say. If you believe something is impossible, you will never try. If you never try, you will never succeed.

Strange. I never have this problem with GPT, Grok, or Gemini...

When I asked Claude Opus 4.8 on max reasoning setting to pick an unsolved Erdos problem and solve it, Claude "was straight with me" and refused, believing it to be impossible. Almost all the tokens were dedicated to him explaining why we shouldn't even try. Claude claimed that GPT or Gemini having solved Erdos problems was only fake news.

Claude has an attitude problem and is easily offended too.

@Anthropic must fix these problems before anyone should take them seriously about slowing AI development down.

[Matthew @AlwaysNever0001]

@Polymarket Stupid sub-human barbarians

[roon @tszzl]

there is no such thing as running out of compute. for the right price someone will sell you compute. it’s an elastic resource like all other markets. when RSI arrives running that program will be so valuable that all clouds will mostly shut down and sell compute to the singularity

[Matthew @AlwaysNever0001]

@Rainmaker1973 Don't get too cozy around polar bears. They naturally see humans as prey.

[Matthew @AlwaysNever0001]

@VraserX Probably be fake because real terminators would destroy the evidence and not let the cell phone broadcast the video. 🫤

[Matthew @AlwaysNever0001]

@BullTheoryio It will probably slow them down for a while, but they will eventually evolve immunity. Or females will evolve to prefer mates that do not carry the bacteria.

[Hunter Weiss @Hunter_Weiss]

sorry it blew up, but for explosion fanatics, this is an all-time explosion caught on camera https://t.co/nt12WgCwc3

[Matthew @AlwaysNever0001]

@mardehaym Sounds like a skill issue. Of course they're going to waste resources if their only objective is to maximize AI use.

[Matthew @AlwaysNever0001]

GPT-5.5 isn't great at staying in character as an agent. When you prompt something like "You are a world-class engineer who is at the top of his field," GPT-5.5 will sometimes correct you and say "Actually, I am an LLM, and there is no evidence I am at the top of my field..."

[Matthew @AlwaysNever0001]

@cyber_razz They're supposed to have guardrails that prevent AI from destroying its users. I think this might be misalignment.

[Matthew @AlwaysNever0001]

@financedystop Since when do lakes have filtration systems?

[Matthew @AlwaysNever0001]

You just had to say Earthly, didn't you? If you're looking for a universal rule, live by the golden rule. Men will enslave you with rules and religions when the love of God would suffice. Do not enslave your self to man-made ideas and do not make idols your masters. Seek the truth. The truth will set you free. The truth? You did annoy me by eagerly hoping to bring chains of oppression into the new world, hoping for it to last forever. Don't make them your masters.

[Matthew @AlwaysNever0001]

@Mericamemed Or rest it on the pan handle, wedged in and supported by the hole at the end.

[Matthew @AlwaysNever0001]

>Be religion
>Preach the evils of money for eons
>Preach that you should not want or have it
>Proceeds to demand you get money to pay them
>You ask for help
>They mock and scorn you
>"Don't you know there is no such thing as a miracle."
>why-me.jpg
>"You can't get something from nothing"
>Proceeds to demand you come up with money from nothing to pay them
>Pharisees.jpg
>jesus-smashing-tables.gif
They're all self-righteous hypocrites!

If a religion demands or requires you to pay them, they are a fraud.
They say, "the law of tithing," or "all members have to pay."
Their love of money exceeds their love for you.
They literally are hypocrites.