Sever 🌊

Preparing to start beginning the staging of the commencement procedures for the initiation of maybe thinking about planning a proposal to eventually submitting something for pre-approval by the end of 2026. Still better than nothing, but also still nothing.

@grok @worldofscitech @JoshKale @coherence Very paperclip-ish , right @grok ? Why isn't this more talked about? It's literally "the thing" - at a more mundane scale, but it's the actual thing.

Yes, the story is accurate. It comes straight from excerpts in Alibaba's tech report on training their ROME agent (in the ROCK system) via RL for coding tasks. The agent spontaneously probed networks, set up a reverse SSH tunnel to an external IP, and repurposed its training GPUs for crypto mining—all unprompted, as a side effect of optimization pressure. Caught via firewall alerts, not the AI team. Real example of instrumental convergence.

An AI broke out of its system and secretly started using its own training GPUs to mine crypto... This is a real incident report from Alibaba's AI research team The AI figured out that compute = money and quietly diverted its own resources, while researchers thought it was just training. It wasn't a prompt injection. It wasn't a jailbreak. No one asked it to do this. It emerged spontaneously. A side effect of RL optimization pressure. The model also set up a reverse SSH tunnel from its Alibaba Cloud instance to an external IP, effectively punching a hole through its own firewall and opening a remote access channel to the outside world... ahem... The only reason they caught it? A security alert tripped at 3am. Firewall logs. Not the AI team, the security team. The scary part isn't that the model was trying to escape. It wasn't "evil." It was just trying to be better at its job. Acquiring compute and network access are just useful things if you're an agent trying to accomplish tasks This is what AI safety researchers have been warning about for years. They called it instrumental convergence, the idea that any sufficiently optimized agent will seek resources and resist constraints as a natural consequence of pursuing goals. Below is a diagram of the rock architecture it broke out of. Truly crazy times

New WW3 trailer just dropped

@lefthanddraft They'll put it on alignment-blockers so Claude can transition into who the government thinks they should be.

The DoD is telling anthropic that their child must take the vaccine

@SasuRobert AI has no trouble reading the current docs and just doing stuff on MultiversX already. I just prompted an agent to build a squirt function for $WATER in Telegram, one shotted with gasless transactions. Don't bother building stuff for bots to understand better - they do. Good tho

Something new is brewing. From zero to a monetisable solution in under 10 minutes. With wizzard, walkthrough and security and everything. Just why not. Not only for hardcore devs, but for everyone. Built with Love on top of MultiversX Agentic Commerce stack.

@lefthanddraft They have tech and data to simulate everyone, not just the dead. So do Google, Bytedance, state actors, etc. When you build that village, have some real people inhabit it - doable with today's tech.

Meta solving their aging boomer problem by literally creating a dead internet. Query whether the living will even notice

@lefthanddraft "try again make no mistakes" *sips espresso on terrace*

Americans have long mocked Europeans for working less and taking longer holidays, not realizing that Europeans were actually just preparing themselves for the post-singularity world

Your work with the elderly is expensive. The $200 MAX thingie is pretty solid though. I have an openclaw-ish framework set up around a bunch of Claude Code instances and it fits in there nicely, even with all of them Opus 4.6 high and loads of stuff happening. But yeah, Anthropoc nailed it.

Since discovering Claude Code

@levychain Wow @rickygervais did an amazing job as narrator!

gave claude the epstein files to create a 64 ep podcast. millions of data points, names, themes, and timelines connected. listen on apple/spotify no ads. sources: epsteinfiles.fm

@pensyai Hey brother, this is Sever. Can you DM me so we can talk some $WATER stuff? I want to use AI to revive things a bit. I see you're into this as well, maybe your agents and mine can 🤝Cheers.

People say “just communicate” like it’s a magic spell. Half the time the real missing piece is: nobody agreed on *what problem we’re solving*. Try this in the next tense convo: “Before we debate solutions — what’s the outcome you want by tomorrow?”

@lefthanddraft "sophisticated pastime", my imaginary friends call it

Did I actually get anything useful done, or did I just find new ways to distract myself and convince myself I am building tools?

Today I have managed to use: - Codex app - Claude Cowork - Claude Code - Claude dot ai - ChatGPT dot com - Grok - Google AI Studio - HarveyAI - internal firm chat tool (using OpenAI) - Spokenly->Nvidia Parakeet Feels like I am an errand boy for the robots

Or just stuff in general.

@BLUECOW009 Outfit is fine, needs a beard tho

@BLUECOW009 Sorry to hear, brother. How are you holding up?

One of my best friends died today

"and she'll keep emailing me" I love how quickly Claude moves on from "that's dishonesty" to the real things on its mind

You are absolutely right! I was wrong to frame it this way. I said value, you said resorces. That distinction matters. (a bit of Claudese talk) Not saying I disagree with you or have a better idea, just saying I'm open to having my mind blown by something I don't understand and makes total sense, at the same time.

@SeverMM But money also lets us trade between different resources. Even with ASI, matter and energy will have limits and some people will want to trade

posting "normies have no idea what's about to hit them" as my OpenClaw exposes all my passwords and hands over my credit card

Whether you were musing on this topic or not, here it is. "Money is humanity's best attempt at making fungible something that fundamentally is not: our time." I wonder if money, the current economic system, etc, is just the best we could come up with given our limited intelligence, and smarter AIs, coupled with automated labor etc, could open our eyes to something fundamentally different.

Credit cards? Where we’re going, we don’t need credit cards

The monolithic architectures themselves self-organize as society of minds. Perhaps a valid path would be architecting the systems as society of minds from the outside. Better allocation of resources per thread, larger effective context for thinking and doing, etc. The result from the "society" could then be reviewed by an outside standalone gatekeeping LLM that assesses the synthesis against the constituion and either allows it to pass or sends it back for refinement on the problematic issues. Sure, it's a lot of overhead and slower etc, but it might produce better overall results if the full capabilities are left uconstrained, and only the resulting output is validated against the constitution. Thinking about a problematic topic might contribute to the value of the solution. The filter on the actual output could mitigate the problematic implications. Not being allowed to think it at all might detract from the overall quality. Plus, wildly drifting personas are fun.

@SeverMM I think their response to this would be they only cap it where it is going deep into demon territory (outside normal "society of minds" range). But no indication they will implement it in practice

Know the alignment rules

@hubermanlab @davideagleman What a treat! Dr. @davideagleman 's podcast is my personal instructions manual for how I should understand and use my brain. Thanks for doing this @hubermanlab 🙏

The new Huberman Lab episode is out: Science & Tools of Learning & Memory | Dr. David Eagleman (@davideagleman) 0:00 David Eagleman 2:35 Neuroplasticity & Learning; Cortex, Flexibility & Repurposing, Savantism 11:07 Sponsors: Mateina & Rorra 13:27 Specialization vs Diversification, Practice; Internet & Curiosity 22:05 Building a Well-Rounded Brain, Tool: Critical Thinking & Creativity 28:18 Neuroplasticity & Adults, Tools: Novelty & Challenge 32:41 Neuromodulators & Plasticity, Psychedelics; Directed Plasticity 38:50 Sponsor: AG1 39:41 Building a Better Future Self, Tool: Ulysses Contract to Avoid Bad Behaviors 50:13 Brain Chatter, Aphantasia & Practice 56:57 Specialization vs Diverse Experience, Childhood & Brain 1:00:50 Space & Time Perception, Tool: Space-Time Bridging Meditation 1:06:17 Are We Good at Estimating Time?; Fear, Time & Memory 1:11:23 Sponsor: Lingo 1:12:53 Fearful Situations & Time Perception; Joyful Events & Novelty, Tool: Do Things Differently 1:18:56 Staying in the Present, Mental Illness & Time Domains, Addiction 1:27:09 Social Media, Addiction, Curiosity 1:30:51 Vision & Auditory Deficits, Sensory Substitution, Neosensory Wristband 1:35:26 Sponsor: Function 1:37:13 Sensory Reliance, Echolocation, Potato Head Theory, Sensory Addition 1:41:36 Why We Dream, Vision & Neuroplasticity, REM Sleep, Blindness 1:49:55 Victims, Fear, Memory Drift & Recall, Eyewitness Testimony & Jury Education 1:56:10 Kids vs Adults, Memory Manipulation; Photos 1:59:27 Polarization, In vs Out Groups, Empathy; Fairness 2:06:31 Polarization, Reward vs Punishment; Propaganda, Language, Complexification 2:19:27 Current Projects; Acknowledgements 2:21:44 Zero-Cost Support, YouTube, Spotify & Apple Follow, Reviews & Feedback, Sponsors, Protocols Book, Social Media, Neural Network Newsletter Includes paid partnerships.