Pinned Tweet
Earlier today I had a baby in one hand and a schooner in the other. Felt like I could rule the world.
dan nolan retweeted
Anthropic marketing: We built a god only we can control and gave it a “wet lab”
Muse marketing: Heyy girl I know you have dozens of unpaid parking tickets. Let’s get that cleaned up
dan nolan retweeted
Disability/chronic illness groups have always been beset with power tripping moral absolutists who treat intersectionality as a points-based hierarchy. In one group I saw a heated argument over whether it's ableist to say that cutting your toe nails on public transport is gross.
dan nolan retweeted
My latest for @TheEconomist: AI writing is taking over politics.
10% of what was is said in Parliament is now AI-drafted, per Pangram. Yet more in the House of Representatives, or in Australia and Canada.
Link: economist.com/britain/2026/0…
And a 🧵 below, w/ some main offenders
It’d rock if you could do something about this
Do tougher job-search requirements help people find work?
We randomised almost 40,000 jobseekers to different requirements. Six months later, employment outcomes were virtually identical.
More in my Shann Lecture on building a learning state. ministers.treasury.gov.au/mi… #randomistas #auspol
Cyber crime is already illegal what do you want man - what ai alignment needs is Australian taxpayers to get involved
shocked so many people are falling for the government’s line that an OpenAI model “hacked” Medicare. I’ve worked in government IT, I’ve literally reported serious vulnerabilities and been shrugged at. I am 99.99% certain the model probably just stumbled across unprotected files.
dan nolan retweeted
Claude? ChatGPT? Singularity? You must have hit your head pretty hard. Hurry up, we're gonna be late for the Bored Ape NFT rave to celebrate Ethereum hitting 4k
If you’re an Aussie into tech and policy stuff I cannot recommend enough @CapitalBrief’s on the call podcast. Short. Sharp. Good great. Thank you @jmcduling - amazing stuff
dan nolan retweeted
On one hand, this advisory simply calls out that AI is at a point where it can perform semi-autonomous vulnerability discovery and exploitation. For those on the coalface, this isn't new information.
On the other hand: It's pretty big deal to see this coming from @asdgovau and @acsc. They're saying that yes, as industry, we should be paying attention to how the economics of attack are in the process of shifting, and that the risks are now real and imminent.
Risks of AI misalignment to Australian organisations | Cyber.gov.au m.cje.io/47dOfOr
