Really, anything.

Everybody with a college course in Thermodynamics.

If you need a plain scope, the Rigol entry-level ones are all decent. But if specifically looking at serial/CAN, may want to take a look at a Saleae-8. The device is headless and comes with probes. Connects to a computer via USB. The software runs on the computer and is what sets them apart.

If too expensive, suggest finding a used 8 and not succumb to cheapo knockoffs on EBay. The software gets regularly updated and runs only on their devices. That alone is worth the investment. You can even extend it by writing your own analyzer code to handle event sequences.

No relation. Just a happy long-time owner of a Saleae-16. Has saved my bacon a few times. Also have an old DS1054, but haven’t had to use it for eons.

A good hockey player plays where the puck is. A great hockey player plays where the puck is going to be.

– Wayne Gretzky.

Wondered where that branch of the family went.

The problem with CAG is not just that it hogs memory, but to keep it fresh you have to keep re-indexing. If the corpus is large and dynamic, it can easily fall out of date and, at runtime, blow out the context window.

GraphRAG has some promise. NVidia has a playbook for converting text into a knowledge graph: https://build.nvidia.com/spark/txt2kg

It’ll probably have the same issues with reindexing, but that will be a common problem, until someone comes up with better incremental training/indexing.

Looks interesting. Will give it a whirl on my home server.

In this article, they talk about bringing up a local RAG system to let people run an LLM off a large document corpus: https://en.andros.dev/blog/aa31d744/from-zero-to-a-rag-system-successes-and-failures/

Wonder if this, connected to something like that, and wrapped in an easy end-user friendly script or UI could be a good combination for a local, domain-specific, grounded knowledge-base?

There was a cheapo Japanese restaurant downtown. Plastic everything. Went there for lunch a while back. Worst Bento box ever.

Six months later. Hmm, Bento box sounds good. Go to this Japanese restaurant. Halfway through the awful meal, remember I’d been there! Swore never to go back. Again.

This cycle repeated SIX times.

What broke it was the whole building burning to the ground because of a grease fire.

Point is… hmm… Bento for lunch sounds good.

They have these on a college campus nearby. I once watched one of these things try to get from a walkway onto a road after a delivery.

It didn’t just backtrack the way it came. It was looking for a sloping ramp, but the sidewalk ledge was too high for its tiny wheels. It kept going back and forth, retracing the same paths over and over, failing, then returning to the original location, turning around and doing the whole thing again. I stopped paying attention after 15 minutes of this.

It was like watching a drunk come out of a bar and stumble around trying to find his parked car.

Gave up after too many two-minute unskippable ad breaks. Also, every streaming service should just use regular announcers from the home team.

She articulates the scientific challenges very well. I would happily subscribe to her podcast/YT/Nebula channel.

There are a TON of third-party attachments online, designed not to cover the camera and enhance privacy, but to cover the little light that reveals it’s live-recording or streaming.

Finding the Money. Worth a watch: https://findingmoneyfilm.com/

Some?

Opportunity for some excellent, weirdo hijinks.

Fascinating talk.

According to the U.S. copyright office and Library of Congress, copyrighted works require a ‘human’ element: https://www.congress.gov/crs-product/LSB10922

If art generated by AI can not be copyrighted, it may well extend to AI-generated code. If so, the implications could be pretty far-reaching.

The one, practical use-case of AI that has found ‘product market fit’ so far has been using AI for coding. Companies are encouraging it. Developers (including experienced ones) are starting to use more of it. But if it turns out none of the generated output can be copyrighted, then you lose all the commercial users who are the revenue sources for all these tools and companies.

This talk feels like it’s touching on a pretty important topic.

uv ---- fork —> wx

(although zx is easier to type)

We are being gaslit. From the article:

Recommended setup: 4x NVIDIA HGX H100, 4x NVIDIA HGX H200, or 2x NVIDIA DGX B200 for optimal performance.

No big. Your typical homelab setup. 🙄

Also: https://github.com/jahrulnr/esp32-picoTTS

Pepsi commercial on side of plane? Who is the target audience? Air traffic controllers and ground crew?