Notes to readers of this draft

Diverse things appear here as fuel for discussion, but many of them may be removed before publishing a final version at the end of each month.

Many of the things in “Threads” have a connection with something being done by at least one participant in this ∮forge.

“xxx” is a placeholder for a link

The Spirits of America xxx by Jeffrey Tucker

How We Invented Freedom & Why It Matters xxx by Daniel Hannan

Born of Ideas: How Principles, Faith, and Courage Forged America x.com by Lawrence W. Reed

[] xxx Reader’s recommendations, as published in the book

Jeffrey Tucker, interviewed xxx by James Howard Kunstler

Joyce Brand’s weekly news from the world of free cities

American Precision Museum

At about 9:45 into: (1) The Equation That Changed Precision Machining Forever - YouTube

Barth’s slide rules, at 11:30

Masters of Production, Borth

A tiny consumer choice became a vote . China . Pears

‘People’s trust in the system has totally collapsed.’

Apple & local AI — https://www.youtube.com/watch?v=1lO8aNSLPJc

‘Apple rebuilt the entire desktop Mac line around local AI, and the price ladder tells you exactly who it is for. Here is what the M6 and M5 chip split means for running local models, keeping agents alive on your desk, and deciding how much AI compute is worth owning.’ —Nate B. Jones

Kiraa.ai, Errol Brandt — (interesting ideas in this, value, to me, unknown. -ls)

https://www.youtube.com/@kiraa_ai

https://www.linkedin.com/company/kiraa-ai/

‘Kiraa is the Knowledge Infrastructure that Powers Enterprise AI. Why? Because most businesses are drowning in data, but starving for insight. Systems don’t talk to each other, information lives in silos, and AI models have no context about how the business actually works. When knowledge is disconnected and scattered, how on earth would you expect AI to make sense? While consumer-grade is helpful for personal things, it's not good enough to make high value business decisions. To do that, you need Enterprise AI. Kiraa delivers enterprise AI by building the missing knowledge layer, the layer that connects your data, your people, and your decisions. With a unified analytical dataframe, shared definitions, and real-time reasoning tools, Kiraa creates the foundation that makes AI useful, trustworthy, and fast inside a business.’ —on LinkedIn

“Apple Just Killed AI Data Centers” — Errol Brandt, CEO of Kiraa

https://www.youtube.com/watch?v=UBArQl_KVzo

‘Apple already has the hard part: extraordinary hardware with unified memory, fast compiled Swift, and a real path to running models locally. What it doesn't have is the enterprise engine — the unglamorous plumbing of job queues, batching, monitoring, auditing, failover and telemetry that turns "I have a model" into "I have an AI system a business can actually rely on." Most developers never build it, because they don't need to. So they run a Formula One car to the shops in first gear — Apple hardware throttled by Python, the GPU idle, the unified-memory advantage thrown away. That missing layer is what we've spent three years building at Kiraa.’

“Redefine AI: Augmented Intelligence” ••• [the video is for paid members. I’m not one of them, so all I can do is post Brandt’s description. -ls]

‘Big tech keeps telling us that AI agents will take over jobs, and for boardrooms chasing shareholder returns, fewer humans sounds like a great deal. That's why we're seeing customer service agents rolled out everywhere, and why cash-strapped government departments are now looking to automate decisions too. But before we hand over the keys, it's worth remembering that we've already tried letting the machine decide, and the results were catastrophic.

In this video I tell the story of Robodebt, the Australian government's automated debt recovery scheme that ran from 2015 to 2019. A simple algorithm averaged annual tax income across fortnights, compared it to welfare records, and issued hundreds of thousands of legally enforceable debts to some of the most vulnerable people in the country, with no public servant ever reviewing the evidence. The burden of proof fell on the citizen, lives were lost, and it took a Royal Commission, a $2.4 billion settlement covering more than 400,000 people, and referrals to the National Anti-Corruption Commission to unwind it. The commissioner called it "venality, incompetence and cowardice."

Robodebt happened before ChatGPT, but the same logic is now playing out inside corporations: remove human discretion, hide behind the algorithm, and make sure nobody's name is on the decision. As someone who builds AI systems, I argue that we should absolutely chase efficiency, but never by removing accountability.

AI should mean Augmented Intelligence: humans making better decisions faster, with a person in charge and a name against every outcome. That's the only version of the future worth building.’