AI: what people are doing and what people are saying (June, 2026)

When it comes to AI, there are 1) lots of people doing practical things with AI and 2) lots of people writing about AI. They rarely overlap. 🙂

But overlap they do in this post! For those of you not doing practical things with AI, skip to the second section.

1) Here’s a bunch of links for people who are doing practical things with AI:

Much is written about models these days, which makes sense since it is a rapidly changing part of technical AI. For example:
- I like this since I run LLMs on my Mac: Best Local LLMs for Mac in 2026 — M1, M2, M3, M4 Tested | InsiderLLM – insiderllm.com
- I use Ollama to run my models, so I liked this too: Top 10 Ollama Models: What They’re Good For and What You Need to Run Them
- This is a big development: 1M context is now generally available for Opus 4.6 and Sonnet 4.6 | Claude. It’s amazing how much a big context window helps.
- This was fun: Introducing talkie: a 13B vintage language model from 1930. Ok not all that practical, but …fun!
- What does the great computer scientist Knuth think of Claude? That paper he wrote will tell you.
- A good piece on different models when it comes to writing: The Creative Gap: Models Split on Writing – makerpulse.ai
- Really there is so much being written about models. For instance:
On IBM Bob, a tool I use a lot: IBM Gets Bob 1.0 Off The Ground
A good piece on how COBOL and AI‘s use with it:
- Can Claude Code Really Understand COBOL Applications?
- Relatedly, IBM weighs in on the topic: Lost in Translation: What the AI code debate keeps getting wrong
Agentic AI is the hot space and has much focus, so here are some agentic AI pieces:
Finally, Rabbit R1’s OpenClaw Update Could Be Its Most Important Moment Yet – Yanko Design

2) Here’s a bunch of links for people who like to read about AI:

On AI and jobs, so much is being written. Take a look:

Some pieces on OpenAI:
How will OpenAI compete? — Benedict Evans – www.ben-evans.com
- - OpenAI is huge in India. Its models are steeped in caste bias.
  - OpenAI Is Making the Mistakes Facebook Made. I Quit.

Finally,
- I thought this is well done: The Star tried something with AI we had never done before
- Lots of people mocked Marc A for this, and rightly so: Analyzing That Controversial Super Prompt For Controlling AI That Billionaire Marc Andreessen Posted
- AI is giving security folks lots of reasons to lose sleep: OpenClaw gives users yet another reason to be freaked out about security – Ars Technica
- Yeah, probably not a good idea: Amazon is determined to use AI for everything – even when it slows down work
- Also not good: Here’s how Israel is using artificial intelligence to find targets in Gaza and Moltbook was peak AI theater | MIT Technology Review

On running GPT-OSS-20B on my Macbook Pro

I’m been running gpt-oss-20b on my Macbook Pro for some time now. I typically run different LLMs on my Macbook using ollama as the basis for this. While I’ve tried various models, this model from OpenAI is the one I found that gave me the best answers to the questions I’ve been giving it. I even prefer the answers I get with gpt-oss-20b on my Mac over those I get from larger SaaS models provided by OpenAI, Anthropic and Google.

I’ve found it very good when it comes to asking it questions concerning lifestyle changes. For instance, I asked it for:

ideas on how to celebrate a big birthday
a 12 month plan to do Swedish death cleaning on my house
a 12 month plan to get my estate in order
a 12 month plan to get back in shape after a long layoff

I’ve found it not very good at all with questions around food and music:

coming up with music playlists based on certain criteria (over a third of the songs were hallucinations!)
coming up with spice mixes for cooking meats (the proportions were way off compared to tested recipes)

Like all LLMs, I found it works well as a means to generate ideas. It works ok or poorly when you start expecting exactitude in the answers. YMMV.

Other things I like about gpt-oss-20b running on ollama on my Macbook:

I can see it go through it’s reasoning when it comes up with the answer, which is a nice feature.
For some reason the default output format is one to my liking, and I can easily move it into Microsoft Word or Apple Pages for further formatting
I especially like it because I can ask personal questions and not have those being captured by a service in the cloud. It provides an extra layer of privacy.

I am running gpt-oss-20b and not gpt-oss-120b due to the capacity of my Macbook Pro. Indeed, even gpt-oss-20b takes up a fair amount of resources on my Macbook, despite it having a GPU to do the heavy lifting. I am not in a rush for these answers though, so it is fine. I’d run 120b if I could, but I have been happy with the response with 20b.

If you have a machine with a GPU and enough RAM and disk space, consider downloading ollama and this model and give it a go. Let me know what you discover!

From RAG to riches (What I find interesting in AI, Dec. 2024)

Here’s my latest post on AI with 60+ worthwhile links found since my last post on it in June. It comes in three sections: the first section is about AI technology, the middle section is on AI in the news and the oped pages, and at the end there’s a good but odd bunch worth a look. Enjoy!

Technology – LLMs:

This is a good repo on tuning Llama2.
This talks about the LLM llama3.
Another good repo: Phi 3CookBook here. More on phi3 here. Also here.
This repo deals with neo4J and LLMs.
Not all language models are small. For example, Reader-LM is a small language model.
Here’s something on LLM pre training and post training.
This here repo is all about building an LLM from scratch. Really good material.
Last, here’s a guide to build your own RAG app: a step by step guide to setup LLMs locally using Ollama, python and chromadb. Nice.

Technology – Ollama / Flowise:

If you need to know how to run ollama in a docker environment, this is your complete guide.
For people wanting to get started with Flowise.
Advice on how to integrate google for flowise.
A piece on Flowise and Langchain.

Technology – RAG:

A simple repo on how to put together RAG and Ollama.
On moving from from RAG to agent systems.
Here’s a recipe for RAG – how cloud services enable generative ai outcomes across industries.
For haters of RAG: in defence of RAG in the era of long-context language models.
Here’s how to harness rag for text, tables and images
If you want master RAG and select an embedding model.

Technology – watsonx.ai:

First off, a watsonx.ai quick start guide.
Two good piece on setting up watsonx .ai to work with python, here and here. Relatedly, this is also very good.
On setting up of watsonx.ai in the cloud.
Go here for the watsonx.ai Prompt Lab.
Here’s info on watsonx.ai’s foundation models.

AI in the news / oped sections:

Looks like Mcdonalds had to end their AI drive thru test. Ah well.
Speaking of AI fail: Detroit facial recognition software had led to false arrests.
More AI fail: Jenna Ortega quit Twitter due to fake AI images being generated of her. Sad.
Here’s a piece in the New York Times on AI chatbots and language learning models. The Times also wrote that the AI boom has an unlikely early winner :wonky consultants.
There’s been a lot of focus of how AI and new data centers are going to cause climate change issues due to a rise in carbon emissions. Relatedly, here’s why AI and clean energy need each other.
On how AI is being used to catch tax fraud.
On improvements of ChatGPT in math and science.
A fun article by Kash Hill who let generative ai make all the decisions in her life.
A piece on how the Quebec government is using AI to track trees, pools, etc, here.
Looks like the “ignore all previous instructions” loophole in chatgpt gpt 4o is being closed. Good.
How AI blew up nanowrimo.
Oops! Google AI is telling people to put glue in pizza. Oh dear.
Here’s all of Dave Karpf’s critiques of IT and the industry. Smart stuff.
Sad stuff: AI took their jobs. Now their job is to make AI sound more human.
On the AI bubble, here.
On AI agents.
On AI slop.

Last but not least:

This is a really good repo / course on generative AI for beginners. (and check out these other repositories here, too).
A good intro to embedding, here.
A tool I really like: autogen studio.
How to create an Azure AI resource.
If you need a ChatGPT cheat sheet, go here.
A good list of prompt engineering tools
Here’s what you need if you want to understand deep learning
On AIops vs Mlops.
This here on the importance of responsible AI.
A step by step guide to using generative AI in your business, here.
A helpful piece on dealing with AI hype, here.

M	T	W	T	F	S	S
1	2	3	4	5	6	7
8	9	10	11	12	13	14
15	16	17	18	19	20	21
22	23	24	25	26	27	28
29	30