Restrainingorder732

[Restrainingorder732 @apexvoidus9CoH]

@INCKerala @elonmusk

[Restrainingorder732 @apexvoidus9CoH]

@waltermasterson Let him play stick games in the muddy rain with his wife

[Restrainingorder732 @apexvoidus9CoH]

@Haleeeemahh @AfnadX In their “selected benchmaxxed benchmarks”
Gemini always has been high scoring in chosen benchmarks and unreliable in practice.
If it is so dangerous in cyber, what is its ExploitBench score?
Keeping defending vaporware

[Restrainingorder732 @apexvoidus9CoH]

@CristiVlad25 @OpenAI It's been terribly worse today and for getting blocked for just opening and analyzing my old submitted report

[Restrainingorder732 @apexvoidus9CoH]

@kimmonismus Dystopian

[Restrainingorder732 @apexvoidus9CoH]

@SadAlbert10 Seeing Google employees hype and cringe post AI-written HR speak while being so disconnected from the reality of their public perception is infuriating. They keep putting their benchmaxxed numbers in front of everybody’s faces. Where is the model?

[Restrainingorder732 @apexvoidus9CoH]

@SThinkTnereffid @elonmusk You are probably a natural born slave, but not all people love to kick boots

[Restrainingorder732 @apexvoidus9CoH]

@Ananth7e 100 dollars to try a super benchmaxxed filtered not relevant model that gets shit on by models on 20 dollar plans by competitors

[Restrainingorder732 @apexvoidus9CoH]

@Seltaa_ Astra is costly, the more odd thing is why not 6.1 Sol?

[Restrainingorder732 @apexvoidus9CoH]

@elonmusk W tweet!

[Restrainingorder732 @apexvoidus9CoH]

@DaveShapi Obnoxiously retarded

[Restrainingorder732 @apexvoidus9CoH]

@bcherny Prompt engineers are copers and just a window for people who have no other skills. It was evident as models got smarter you don't need to throw up the third known gang sign from mesopotamia to unlock a "hidden reasoning model"

[Boris Cherny @bcherny]

I am surprised that people are surprised this is how I prompt Claude.

Talk to Claude the way you would a coworker. There's no secret to prompting. There's no need to be overly scaffolded or prescriptive for most tasks -- give Claude a goal, and it will figure it out.

Back in the Sonnet 3.5 days, your prompt mattered a lot. Nowadays, it's much more important to communicate to the model:

1. What you want it to do
2. How much effort you want it to spend
3. How it should verify that it did the right thing

[Restrainingorder732 @apexvoidus9CoH]

@keikane_ Is the red in the room with us?

[Restrainingorder732 @apexvoidus9CoH]

@TokenGremlin The last day was still useful arguably

Today feels like meh

[Restrainingorder732 @apexvoidus9CoH]

@Chaos2Cured @AnthropicAI I am curious, when you say the labs "hate you*
What happened that led you to this conclusion?
Not defending the labs btw

[Restrainingorder732 @apexvoidus9CoH]

@Chaos2Cured @AnthropicAI I don't think it works like that, the internet isn't clash royal where you drop a level 8 pekka, your opponent does too and both cancel each other out. Offense is always at an advantage in cybersecurity

[Restrainingorder732 @apexvoidus9CoH]

@Chaos2Cured @AnthropicAI Yeah but tbh that would also increase chances of other script kiddies with similar tools attacking you or anybody.
I mean I am all for normal AIs to be shared broadly amongst people, concentrating that to few would cause a bad power imbalance but this is inherently risky

[Restrainingorder732 @apexvoidus9CoH]

@Chaos2Cured @AnthropicAI I mean not everybody needs cyber models😭😭😭😭

[Restrainingorder732 @apexvoidus9CoH]

@Seltaa_ Grok 3 and 4.1