Mapping Billions of Dollars of Data Center Capital Expenditure An Entity Resolved Semiconductor Business Graph

@wildfranz.bsky.social @niamhmcintyre.bsky.social @tbij.bsky.social Think you might be interested in this from @rjurney.bsky.social

blog.graphlet.ai/mapping-bill...

I’m supposed to come here for a nicer Twitter, but the Mac app can’t get larger than an iPhone?

I am amazed at how helpful @notebooklm.bsky.social is for writing historical fiction… I know it’s just RAG, but it’s good RAG you can ask questions about what this or that would be like in your story!

What was it?

Actually I found out NordVPN is keeping MeshNet due to outcry from their users :)

On the possible use of hydraulic force to assist with building the step pyramid of saqqara The Step Pyramid of Djoser in Saqqara, Egypt, is considered the oldest of the seven monumental pyramids built about 4,500 years ago. From transdisciplinary analysis, it was discovered that a hydraulic...

I am gonna try to hit y’all with ancient history on a daily basis.

Know that ancient aliens shit? Humans own those off-world fucks. Djoser’s Stepped Pyramid used hydraulics to shit 80 ton blocks out the top like a volcano 4,650 years ago… hydraulics still exists.

journals.plos.org/plosone/arti...

Does anyone know of an alternative to @nordvpn.bsky.social Meshnet? I *love* this product and must find someone else by November 1st... five minute VPN ftw!

cc @decius.bsky.social you knows the security, right?

My Experiments with NotebookLM for Teaching  | Towards Data Science Exploring NotebookLM as a teaching companion

Transform traditional study materials into interactive games with NotebookLM and CrosswordLabs. Parul Pandey's new article dives into how to engage students using technology to make lessons more enjoyable and memorable.

The Rise of Semantic Entity Resolution | Towards Data Science Semantic entity resolution uses language models to bring an increased level of automation to schema alignment, blocking (grouping records into smaller, efficient blocks for all-pairs comparison at qua...

The Rise of Semantic Entity Resolution is on Towards Data Science this morning…!

> semantic entity resolution for knowledge graphs uses language models to automate the most painful part of building knowledge graphs from text: deduplicating records.

towardsdatascience.com/the-rise-of-...

This is really funny… arbitrary bitcoin KYC is far worse than fiat banking. Make one questionable purchase ever in your life… account DENIED. No appeal. Regulation of finance does not suck. Censorship for LIFE.

from gremllm import Gremllm # Be sure to tell your gremllm what sort of thing it is counter = Gremllm('counter') counter.value = 5 counter.increment() print(counter.value) # 6? print(counter.to_roman_numerals()) # VI?

The is diabolical... a Python object that hallucinates method implementations on demand any time you call them, using my LLM Python library github.com/awwaiid/grem...

In that same period Google and Hadoop birthed big data, which formed the basis for the current AI trend. There were huge developments like MaReduce.

Well, LLMs are better at teaching math and code than any other method… if you don’t get it, just ask a lot of questions to an LLM with grounding so you can verify on the source.

Real utility from… NFTs?

Okay, domains is a reasonable use case. It is the first one with real utility I have seen. I don’t think removing fungibility from media is a good thing.

Maybe I missed it… what purpose do NFTs serve?

LLMs are generative AI, in the field of artificial intelligence, in the field of computer science. They are rapidly transforming our world. NFTs did what for society? So far only cryptocurrency has impactful utility.

My `CLAUDE.md`, command and allows-tools setup - note that everything has a CLI :) Why not? My `CLAUDE.md`, command and allows-tools setup - note that everything has a CLI :) Why not? - CLAUDE.md

Check out what productive settings for a Claude Code Python project looks like… CLAUDE.md, commands and allowed-tools. This setup makes me 5-10x as productive as before AI.

Feedback appreciated!

gist.github.com/rjurney/4d6c...

My `CLAUDE.md`, command and allows-tools setup - note that everything has a CLI :) Why not? My `CLAUDE.md`, command and allows-tools setup - note that everything has a CLI :) Why not? - CLAUDE.md

It upsets me honestly, because I know they’re getting laid off while early adopters of AI coding will get ahead. It is a _skill_ to get Claude Code to work for you on your project. AI still needs a human, albeit at a higher level.

Here’s what I mean… this took time:

gist.github.com/rjurney/4d6c...

For my next trick, I’ll short TSLA for the sixth time in a few months. Made a bunch on four, lost a little on the fifth :)

I will admit that I do all features first as CLI features as a PM strategy enabled by agent coding and… it never gets my CLI instructions wrong, so I just glance at them. It has never made a mistake or done something unexpected in fifty commands.

It’s a definite risk, and an emergent property already if you don’t refactor, so some productivity must be lost on making it readable.

Yes we do, you’re still responsible for the code. Read and understand every line. Interrupt the agent. Give it feedback. Constantly.

This is what bad AI generated code is like, actually, and it’s shit I reject from PRs. You have to tell it to get that shit out of here, build a simpler interface and update your project config to adjust the behavior. It’s all a software engineering skill.

Okay, good. Because now it’s 30% at Microsoft.

It’s amazing. I want to help you, but you think the thing putting you out of work is an NFT.

More than 30% at google. And rapidly growing. Both those are top software orgs. AI needs you to operate it, if you don’t learn how you’re a lamplighter.

How much code do you think LLMs are writing at this point, percentage wise? 20% at Microsoft, and that’s LOW. But it’s like NFTs? I’m talking to someone getting laid off in 2026, maybe even 2025, for your attitude re: LLMs.

No, this has happened. Textbook rental companies died. Now stack overflow is dead.

Wait, what? You’re just as responsible for what you commit with agents so… how is there little incentive to ask? Precise definitions are in fact now the entire field of software engineering.