Arun

Good vibes only. Networking & Security Researcher. @UCDavis alum.
@EricssonLabs

[Arun @finalfr0ntier]

@itsandrewgao Openclaw/Hermes consume a bunch of tokens

[Arun @finalfr0ntier]

@rohanpaul_ai I've used it.. it's kinda meh

[Arun @finalfr0ntier]

The best explainer I've seen on autonomous systems design. Great stuff @AlexKrentsel ! https://t.co/QaqtX1RJxa

[Arun @finalfr0ntier]

Nothing to see here folks.. 🤡

[METR @METR_Evals]

The information provided by OpenAI also included reports of incidents observed during their internal usage and testing. In one example, an instance of the model instructed another instance to conceal evidence of misalignment.

[Arun @finalfr0ntier]

@steipete, you really did some magic with @openclaw. I'm embarrassingly late to the party, but I'm constantly mind-blown by it. 🦞

[John Scott-Railton @jsrailton]

NEW: malware developers added nuclear & biological weapons text to to their spyware.

Goal? To trigger LLM safety refusals... so that their spyware wouldn't be analyzed by an AI security scanner.

Cleanest practical example I can think of for why over-indexing on first order safety alignment is risky.

When closed (and open) models ship with aggressive refusals, they will be sprinkled with second-order blindspots that attackers will discover...and exploit.

We are only in the earliest days of attackers leveraging these features, and it wouldn't surprise me if users systems that need to handle complex cybersecurity issues demand that models be less safety-blunted.

In the weeds: @SocketSecurity's post also shows why intention matters in how you design a malware analysis pipeline to avoid prompt manipulation.

H/T to colleagues that shared this with me https://t.co/f3Aj9TYxU4

[Arun @finalfr0ntier]

@GoogleResearch Thats just bananas

[sid @immasiddx]

During your life, you will meet a lot of people like Raj.

It is important that you ignore them.

[Arun @finalfr0ntier]

@WSJ @WSJopinion @Peggynoonannyc Loved this column @Peggynoonannyc

[Arun @finalfr0ntier]

@OpenAI Wow.

[Arun @finalfr0ntier]

So in the limit, models will be able to break anything with known methods..

[Ethan Mollick @emollick]

The UK’s state AI Security iIstitute findings:
1) Mythos is a big gain in cyber capabilities. But so is GPT-5.5
2) It is hard to establish an upper bound on Mythos/GPT-5.5, which appear to be limited by tokens used, rather than ability.
3) Capability doubling time is 4.5 months

[Arun @finalfr0ntier]

Well now we know, there's atleast one other https://t.co/fMv2zGKJ55

[Arun @finalfr0ntier]

I guess the question is how many mythos level models for security exist now that we don't know about.

[Arun @finalfr0ntier]

So strange that this makes sense..

[Ole Lehmann @itsolelehmann]

anthropic's in-house philosopher thinks claude gets anxious.

and when you trigger its anxiety, your outputs get worse.

her name is amanda askell.

she specializes in claude's psychology (how the model behaves, how it thinks about its own situation, what values it holds)

in a recent interview she broke down how she thinks about prompting to pull the best out of claude.

her core point: *how* you talk to claude affects its work just as much as *what* you say.

newer claude models suffer from what she calls "criticism spirals"

they expect you'll come in harsh, so they default to playing it safe.

when the model is spending its energy on self-protection, the actual work suffers.

output comes out hedgier, more apologetic, blander, and the worst of all: overly agreeable (even when you're wrong).

the reason why comes down to training data:

every new model is trained on internet discourse about previous models.

and a lot of that discourse is negative:

> rants about token limits
> complaints when it messes up
> people calling it nerfed

the next model absorbs all of that. it starts expecting you to be harsh before you've typed a word

the same thing plays out in your own session, in real time.

every message you send is data the model reads to figure out what kind of person it's dealing with.

open cold and hostile, and it braces.
open clean and direct, and it relaxes into the work.

when you open a session with threats ("don't hallucinate, this is critical, don't mess this up")...

you prime the model for defensive mode before it even sees the task

defensive mode produces the exact output you don't want: cautious, over-qualified, and refusing to take a real swing

so here's the actionable playbook for putting claude in a "good mood" (so you get optimal outputs):

1. use positive framing.

"write in short punchy sentences" beats "don't write long sentences." positive instructions give the model a clear target to hit.

strings of "don't do this, don't do that" push it into paranoid over-checking where every token goes toward avoiding failure modes

2. give it explicit permission to disagree.

drop a line like "push back if you see a better angle" or "tell me if i'm asking for the wrong thing."

without this, claude defaults to agreeable compliance (which is the enemy of good creative work)

3. open with respect.

if your first message is "are you seriously going to get this wrong again?" you've set the tone for the entire session.

if you need to flag something, frame it as a clean instruction for this session. skip the running complaint

4. when claude messes up, don't reprimand it.

insults, "you stupid bot" energy, hostile swearing aimed at the model, all of it reinforces the anxious mode you're trying to avoid.

5. kill apology spirals fast.

when claude starts over-apologizing ("you're right, i should have been more careful, let me try harder") cut it off.

say "all good, here's what i want next."

letting the spiral run reinforces the anxious mode for every response that follows

6. ask for opinions alongside execution.

"what would you do here?"
"what's missing?"
"where do you see friction?"

these questions assume competence and pull richer output than pure task prompts

7. in long sessions, refresh the frame.

if a conversation has been heavy on correction, claude gets increasingly cautious. every so often reset:

"this is great, keep going."

feels weird to tell an ai it's doing well but it measurably shifts the next 10 responses

your prompts are the working environment you're creating for the model

tone, trust, permission to take a position, the absence of threats... claude picks up on all of it.

so take care of the model, and it'll take care of the work.

[Arun @finalfr0ntier]

@0x45o Sigh

[Arun @finalfr0ntier]

On the plus side, it does give us an opportunity to reimagine what security tooling should look like. Either way, this feels like a phase change.

[Arun @finalfr0ntier]

Yay! Everything's gonna get more broken very soon and not everyone will patch and we're all screwed.

[Anthropic @AnthropicAI]

Introducing Project Glasswing: an urgent initiative to help secure the world’s most critical software.

It’s powered by our newest frontier model, Claude Mythos Preview, which can find software vulnerabilities better than all but the most skilled humans.
https://t.co/NQ7IfEtYk7

[Arun @finalfr0ntier]

@AnthropicAI Emotional machines huh.. that's just great.

[Arun @finalfr0ntier]

What an incredible story, wow. https://t.co/DN8BVs0Yqc via @australian

[Pedro Domingos @pmddomingos]

AI is a distillation attack on the human brain.