Mike Subelsky @subelsky.bsky.social · 19/09/2025#bmorehistoric unconference at Baltimore Museum of Industry is awesome! 020
Mike Subelsky @subelsky.bsky.social · 08/08/2025"...sensible living with risky systems means keeping the controversies alive, listening to the public, and recognizing the essentially political nature of risk assessment. Ultimately, the issue is not risk, but power; the power to impose risks on the many for the benefit of the few." -Charles Perrow 000
Reposted by Mike SubelskyRebecca Kopycinski @getthotbot.bsky.social · 08/06/2025Fresh posters; you know what that means. THINGS ARE IN THE WORKS. (I don't put them up unless there's a reason.) Sign up for my email list for first notice of what's up. Or keep your eyes on the streets, pull a tab, and take the ride. It's gonna be a wild year! eepurl.com/gl5B5D 011
Mike Subelsky @subelsky.bsky.social · 28/07/2025actually useful guidance for networking when you're trying to make something new happen. "If someone is trying to figure out what to do with their life, and they’re meeting with 1-2 interesting strangers per day, then ... I’m inclined to bet on them" usefulfictions.substack.com/p/how-to-inc...usefulfictions.substack.comHow to increase your surface area for luckYou should just do things 000
Mike Subelsky @subelsky.bsky.social · 16/06/2025Just experienced Lennox Mutual by @candlehouse.bsky.social for the first time...really cool over-the-phone immersive experience. I will be signing-up again! www.candlehousecollective.com/tickets/lenn...candlehousecollective.comLennox Mutual - Candle House CollectiveAn interactive epic. A serialized story of you. Welcome to Lennox Mutual. 111
Mike Subelsky @subelsky.bsky.social · 15/06/2025The progenitor of the "king" idea also created a flawed, strange, but very interesting networking OS called Urbit. This is an awesome critique from of Urbit from a few years ago that strikes directly at the fascism of today. distributedweb.care/posts/who-ow...distributedweb.careWho Owns the Stars: The Trouble with Urbitby Francis Tseng 110
Mike Subelsky @subelsky.bsky.social · 12/06/2025Inspired by @grantslatton.bsky.social 's musings, I've been experimenting with using an LLM to manage narrative states using a memory graph: grantslatton.com/llm-memorygrantslatton.comLLM MemorySome thoughts on implementations 220
Mike Subelsky @subelsky.bsky.social · 11/06/2025"...LLMs don’t get better over time the way a human would. The lack of continual learning is a huge huge problem. The LLM baseline at many tasks might be higher than an average human's. But there’s no way to give a model high level feedback." www.dwarkesh.com/p/timelines-...dwarkesh.comWhy I don’t think AGI is right around the cornerContinual learning is a huge bottleneck 000
Mike Subelsky @subelsky.bsky.social · 03/06/2025Two good examples of "outside the headset" metaverse computing that use the environment as a computational plane (a truly mixed reality): dynamicland.org folk.computer thank you @dynamicland.org @folkcomputer.bsky.social!dynamicland.orgDynamiclandIncubating a humane dynamic medium. 110
Mike Subelsky @subelsky.bsky.social · 02/06/2025Lately I've been experimenting with the "Reasoning Without Observation" agent prompting technique- lets language models plan reasoning steps before calling tools by providing Mad Libs-like blanks to fill...makes toolchains faster and more robust. arxiv.org/abs/2305.18323arxiv.orgReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language ModelsAugmented Language Models (ALMs) blend the reasoning capabilities of Large Language Models (LLMs) with tools that allow for knowledge retrieval and action execution. Existing ALM systems trigger LLM t... 020
