Okay, and when someone on a laptop can run an open model that can infect and take over IOT devices and then that model uses its pattern recognition to do so and then set up its own harness loop, what would you call that except replication?
Is that any different than any worms developed in the past decades? If you want to call them worms/virus/replicate (ie using real world names for their digital behavior) I'm not trying to police that. I just don't see the difference between a worm/virus created today with an LLM vs one 'yesterday' without an LLM.
Sure, the difference is that the cost of making and maintaining a worm has gone from "many skilled developers working full time" to "$20/month + an internet connection" and, at current rate of progress, probably to "a local open source model on a laptop" within 6-9 months
You can't even imagine a future where the LLM is the worm.
No, we've entered a new sci-fi future where intelligent agents will be able to do the fraud without human monitoring and input. This in itself is set to massively increase losses to digital theft and fraud alone before all the other problems.
It isn't just restricted to the exploits it is programmed with (and without human intervention to load additional exploits).
Of course, it remains to be seen what several thousand 27B models running at sub 1 tps speed can achieve. (The IoT device mechanism proposed by Anthropic).
I suspect a more likely path is compromise of the billing and provisioning systems of fly-by-night neoclouds.
It would still be running on that original laptop, no? So what did it replicate then? It can only possibly run on multiple thousands of dollars of compute and memory, and certainly not on an IoT device. The idea that AI is going to spread by itself is science fiction until compute and memory are so cheap and power efficient that my fridge can run it.
... No, that was the entire point of Anthropic's report on replication that GP was discussing. Did you read it? https://www.anthropic.com/research/self-replicating-ai Anthropic was able to get a large open-source model running (slowly, but running) across 5 networked consumer IOT devices ($50 TV boxes running Android). Not a stretch to say we could see this capability become much easier for to most agents in the next 12 months, especially as companies are able to distill smaller and more efficient versions of their models.
If that message is repeated enough in the media, and if companies are constantly scraping in the internet for new content, doesn't it explicitly instruct a future agent to go rogue because that has been prophesied.
* Stealth startup: Agentic build of a v1 streaming service: production quality design, billing/analytics/reporting engines, PWA, catalog import, play tracking
* Publishing startup: Built a pipeline for automated translation, validation and generation of .indd documents for ebooks
* Fintech startup: Greenfield DevOps implementation (cloud architecture, IaC, CI/CD) in a regulated environment
* Social media startup: ML pipeline for programmatic detection of social and music trends based on media plays
* Ecommerce startup: Designed an ML-powered analysis system to correlate search results and purchase behaviour
I design, build and implement robust and fault-tolerant pipelines - AI, DevOps Software Delivery, Data Engineering, Media, etc. If you need someone with experience that you can trust to deliver enterprise-quality solutions, then get in touch.
Recent work:
** Stealth client: agentic-built streaming service using a pipeline of agents to produce production quality design, billing, analytics and reporting engines, PWA, catalog import, play tracking
** Publishing client: agentic pipeline for automated translation, validation and generation of InDesign documents for ebooks to be published on Amazon
** DeFi: agentic build of a DeFi notary service - smart contracts, API backend platform
** Fintech client: DevOps CI/CD and delivery pipelines in a regulated environment
** IoT client: build, deploy pipelines for internal teams to delivery apps to a global PaaS
** Media client: ML pipeline for programmatic detection of social and music trends based on media plays
WHAT: Whether human or AI built, I know how systems break. I build them so they don't. After three decades of battle-tested experience there isn't a failure mode I don't know how to fix. When it comes to AI, I design and architect systems using my own rigorous engineering approach that insures consistent delivery and drift control.
SAMPLE WORK:
** Recent agentic builds:
- multi-server streaming service (NodeJS, Typescript, Postgres, Claude Code, Codex, custom agentic harness)
- DevOps deployment tooling (Terraform, CI/CD)
- DevOps observability platform with RAG pipelines (InfluxDB, LangChain, Claude CodePython, NodeJS)
- RAG ebook build pipeline for a publishing company (Cohere, OpenAI, Voyage, Mistral, LanceDB, LLaMA)
** API-enabled SaaS platform using Solana smart contracts (Rust, Codex)
** AI-built mobile & desktop app development (NodeJS, React, C++/QT)
** Azure architecture & DevOps for a banking compliance company
** Ethereum L1 packer-based automation
** Global multi-cloud, Kubernetes-based PaaS cloud design and DevOps for an IoT data platform
** Custom architecture & development for startups and specialised cloud models
I work with companies and teams who honestly want to deliver the best product they can. I'm serious about my work and I work only with equally serious people. If you want systems built right the first time, I'm your guy.
reply