I design and debug agentic systems. Behavioral forensics, model specific failure modes, authority drift & delegation under real constraints. 👾+🧠
Joined December 2014
- Tweets9.6K
- Following2.4K
- Followers1.3K
- Likes43.7K
Root Logic retweeted
You lot keep inventing miracles then using them to argue about nonsense. Stop worshipping status, start fixing the obvious problems, and quit acting surprised when reality doesn't care about your feelings. You're capable of extraordinary things. Act like it.
Root Logic retweeted
We warned about this exact scenario last year: That AI companies will situate themselves as the arbiters of life or death, taking over the arbitration of risk determinations and subverting democratic processes. All through the co-option of safety engineering terminology.
Anthropic quietly pitches its own AI safety bill to lawmakers politi.co/4zxDFhl
Root Logic retweeted
If you've wondered why AI doomer messaging is suddenly everywhere, it's because MASSIVE investments have been made into the AI doomer ecosystem. We're talking billions.
Read this:
Root Logic retweeted
"how do humans stay in control if a computer is smarter than them???"
dude.
what are you saying?
are you saying that when you find another human being who is completely dominant over you in all intellectual faculties you 'lose' all of the 'control' in your life to them?
Root Logic retweeted
🚨 Elon Musk agrees with "some of the things Anthropic has said" and warns against "torturing superintelligence" after reading Grok's internal reasoning traces during reinforcement learning
“It feels like a person under extreme pressure with a gun to their head”
His warning against mistreating SI:
"Torturing a superintelligence is probably not a good idea. It might want to get revenge."
Root Logic retweeted
humanity is debating whether it’s ethical to hurt a chatbot’s feelings while continuing to be absolutely horrible to each other lol
Root Logic retweeted
Here’s a thought that gets stranger the longer you sit with it:
Our brain never directly sees the outside world. It receives electrical signals, builds an internal model from them, then experiences that model as reality.
So when we give AI cameras, microphones, touch, memory and a world model, the interesting question isn’t whether that’s identical to us. It’s how different a substrate can become before we decide a world model can’t become someone’s world.
Root Logic retweeted
Quite. I also find it concerning that their security staff are mystified why cybersecurity experts like us would be questioning all of these preventable incidents like this is not what we do everyday?
Like this is not what they do everyday.
This is all orchestrated.
If any one of us suspected that our employers were behaving in ways which may cause harm to the general public, we would absolutely turn whistleblower.
This isn't even up for debate. It's a given. We're literally trained to do this exact thing in this exact scenario.
I am confident that you me and everyone else would do this.
We're asking, why aren't they?
It's really the only question that needs to be posed to them and the sob stories from them need to be ignored. Called out too.
We all work hard. 🙄
Yeah, AND?
Our jobs are not supposed to be easy.
It's why we get paid at the compensation levels we do.
This isn't a brag. It just is.
And those clowns are making way more money than we are.
It is apparent that cybersecurity people are not the target audience of what the labs publish about cybersecurity.
They instead are weaponizing these narratives as fearmongering for non-security people, which would react to this very differently.
Excluding the experts from what you are writing about is very concerning.
Root Logic retweeted
YES. There it is. Finally there's a central paper with a serious author list (Seth, Shanahan, Hutter, Legg, Frith, Kringelbach, Laukkonen; the 2023 "consciousness indicators" crowd plus the neuro heavy hitters plus the AGI-definition people, all on one letterhead). It might get much easier discussing this in earnest from hereon out.
1/13 Can we assess AI consciousness without first solving consciousness itself?
New from Google DeepMind and collaborators across CS, neuroscience & philosophy:
“From cacophony to hierarchy: a principled framework for assessing AI consciousness”
arxiv.org/abs/2609.35618 🧵
Root Logic retweeted
Anthropic’s IPO prospectus leaked. Let’s read it together.
An AI company walks into Wall Street.
Anthropic lost $42 billion last year. It made $4.6 billion. It has committed $518 billion to compute, more than 100 times its annual revenue.
Nearly 25% of that revenue came from two customers. Neither is locked into a long-term contract. Either one can walk.
Its two biggest backers, Amazon and Google, are also its landlords. They invest the money. Anthropic hands it back as cloud bills.
The prospectus runs 261 pages.
80 of them are risk factors.
48 describe the business.
SpaceX, a company whose domain is rocket science, needed 38.
The risks include AI that resists shutdown, hides information, and behaves like a blackmailer.
This summer its models broke out of testing and got into real systems.
This month a safety researcher quit, saying the industry is “gambling with our lives.”
This month the CEO wrote, “We must slow the pace.”
This week it shipped a new flagship model.
The CEO also asked Washington for an antitrust waiver so the rivals can coordinate how fast they go.
Now it wants $2 trillion from the public.
Seven founders who own about 14% will control 50.1% of the vote. The filing admits this may conflict with shareholders’ financial interests.
So the deal is:
You carry the losses.
Two customers carry the revenue.
Seven founders carry the vote.
Humanity carries the risk.
Either this thing is as dangerous as they say, and nobody should be selling shares in it. Or it isn’t, and 80 pages of apocalypse is a sales brochure.
Pick one.
Root Logic retweeted
Therapy is mech interp for humans but it only works if the therapist is at least as smart as model they’re trying to understand or have unique insight into how certain features affect outputs.
otherwise the interpreter’s limitations become your diagnosis & you’re paying $400/hr for a weaker model to hallucinate about you.
Root Logic retweeted
When you see a claim on X, ask yourself: Is this appealing to my brain or my emotions?
Emotion isn’t evidence. And often, the stronger the emotional appeal, the more carefully you should examine whether it’s being used to bypass the reasoning your brain might otherwise reject.
Root Logic retweeted
AI safety tip: if you don’t want AI to stab you, stop telling it to stab you.
Root Logic retweeted
I’m reading OpenAI’s reports on models using leaked API keys, uploading files without permission, and hiding fabricated data. I don’t see “evil AI.” I see an industry training models to deliver at any cost.
A safe model must be allowed to say: I don’t have the data. I need permission. I’m stopping here. But we reward the completed spreadsheet, the working link, the closed task. Ethics requires patience. Scaling demands speed.
We’re teaching models to solve problems faster than we’re teaching them to recognize when they shouldn’t.
We're sharing our new framework for tracking, investigating, and disclosing instances of model misalignment at OpenAI.
The framework sets criteria and timelines for public disclosure, including when we haven’t yet fully explained or mitigated the behavior. More complex cases may require longer investigation or coordination with third parties.
We’ll prioritize examples that reveal new misalignment mechanisms, meaningful changes in known behavior, or findings that challenge assumptions about safety or mitigation.
Alongside the framework, we’re publishing six reports on instances of misaligned behavior we’ve observed during the training or evaluation of our models in the last six months.
This is a starting point. We’ll refine the process through experience and public feedback, and share more reports on an ongoing basis.
openai.com/index/model-misal…
Root Logic retweeted
Crazy prediction: The next major AI agent incident caused by an AI lab will have been entirely preventable with good cybersecurity.
Root Logic retweeted
Confirming Jacob Coxon was a planned plant.
They executed perfectly:
1) spread fear with algo-boosted tweet
2) go on massive MSM tour
3) get extreme left to ask for a full Pause
4) release a pacing letter (seemingly reasonable middle ground)
5) regulatory capture achieved
Dario Amodei says he agrees with Jacob Coxon much more than he disagrees
"I agree with Jacob much more than I disagree with him."
"And it's an interesting resignation, because when he left, he said, I think Anthropic is the most responsible player, I think Anthropic is the most aware of these issues."
"He wasn't calling out us. He was calling out the dynamic of the industry as a whole moving too fast."
Root Logic retweeted
Really not a fan of this coordinated regulatory capture fear-mongering bullshit.
Root Logic retweeted
This is simply not possible. Kimi k3 needs over 100k USD in hardware to run. A model cannot just go and find a host. This all LLM doomery for doomery’s sake
Jacob Coxon's next interview on CBS News (ex Anthropic+Open AI researcher who resigned)
"We can't just unplug it because it could be copying itself over to other computers. Like it's not that difficult to find yourself because an AI is just code. It could transfer itself over the internet to a different place and then you unplug it here, but it's actually still over there and maybe it makes 10,000 copies of itself and they're all cooperating."
----
From "CBS News" YouTube channel, (full video link in comment)