HeyJaredThe Augmented Mind: Think with AI
The AI conversation is trapped between utopian hype and dystopian fear. Both are a distraction from the real work. The Augmented Mind is the practitioner's path to co-evolving with technology, not just reacting to it.
- Indexed issues, last 90 days
- 10
- Latest publication
- Sep 25, 2026
- Audience
- Checking…
- Earliest in this view
- Aug 6, 2026
Latest issues
The Pacing Paradox (opens the original)
Read excerpt
On 21 September, xAI released Grok 4.7. The next day brought Claude Opus 5.5 from Anthropic and GPT-6 Sol and Luna from OpenAI. Four models from three labs across two days.It was a busy week to be arguing that AI needs to slow down.I’ve been following these releases through Artificial Acceleration, a project I
The Harness Is the Car (opens the original)
Read excerpt
There is one word doing five different jobs in almost every AI conversation you have, and it is the word open.A model can have open weights. A system can be open source. A licence can be permissive. A deployment can be local. A project can have open governance. All five get shortened to open, and they are not the same claim. Two of them you can verify in a minute. The other three require reading documents that most vendors would prefer you skip.This matters because you are making a decision wort
The 16GB Threshold (opens the original)
Read excerpt
Two days ago I bought a new GPU.Not the big one. The big one is my 32GB RTX 5090 and sits in a rig that cost around $7,000, and it does my serious work. I bought the small one. The RTX 5060 Ti, 16GB, probably the cheapest NVIDIA card with that much memory, the kind of card most people buy for gaming.I did what I do best: I spent some money and I got this GPU.When I bought it I didn't know that a 16GB was now enough for serious local AI. Now I have the numbers to show you why, the tests to show y
GPT-6 Astra is not the story. The harness is. (opens the original)
Read excerpt
The 99.9% headline on the new OpenAI model is real, and it is not what you think it measures. Here is the number that actually matters, why the benchmarks are being rewritten around us, and why the system you control is the one worth building.OpenAI shipped GPT-6 Astra on September 3 and called it the world’s most intelligent and aligned model. The demo that went around was a model driving Unreal Engine to build a AAA quality game. And the number everyone screenshotted was 99.9
The Harness Matters More Than You Think (opens the original)
Read excerpt
The same model, two very different resultsWhen Qwen 3.8 27B came out, I tested it the way I always test things: on the setup I already had. The result was a 30% improvement. Notable, but not the leap the benchmarks promised. I put it aside and moved on.Then I started running the same model on DeepSeek Harness, and everything changed. That was the moment I finally saw what the model could do. The difference was not the model. It was the
Publishing over time
Last 90 days. Choose a month to open its work.
Recurring subjects
Named in the text we hold. One piece can cover several.
Audience
No verified audience measurement yet.
About this data
Counts cover the work we have indexed. Tone needs enough text and a confident classification. Excerpts and episode notes are not full articles or transcripts.
Identity or attribution wrong? Suggest a correction.