aldo
612 posts


@BrianRoemmele @grok Hey Brian what do you make of the mempalace seems like a way to search offline history
English

How to not pay for an AI agent.
Get a computer you don’t use. NO NOT AN APPLE MAC STUDIO (like and subscribe).
Use something far less expensive, ask @Grok to find a useful pile of junk on eBay.
And here are the complicated steps:



English

The self-proclaimed Mother of All Games! Did you play this back in the day with your friends?
Scorched Earth, by Wendell Hicken, was released as shareware in 1991. It wasn't going to win awards for groundbreaking graphics, yet it's one of the best multiplayer games ever - highly addictive with endless replay value.
It let you customize a huge range of settings, from gravity, wind, and meteor showers to many others. Building on earlier games like QBasic Gorillas, it took the concept to a much higher level, supporting up to 10 players with far greater complexity and variety.
You set angle and power to aim, with a wide choice of weapons (unlike the banana in QBasic Gorillas). Computer-controlled enemies ranged from difficulty 1 (Moron) to 10 (Cyborg). No other game has named its lowest difficulty level so perfectly!
Wendell Hicken, the Quentin Tarantino of game developers!
English

@BenjDicken So creating an SGA like oracle, should als do direct I/O and bypass OS cache
English

Did you know Postgres uses TWO in-memory caches?
Every time pg reads a row, it comes from one of three places:
1) Postgres-managed shared_buffers cache. This is the "closest" layer of disk pages to Postgres, fulfilling the fastest.
2) OS page cache. Operating systems utilize unused memory as a cache of recently-read pages from the filesystem. A Pg read can miss on shared_buffers, but the page exists in the OS cache. This is still a very fast read, just need to copy into shared_buffers.
3) Filesystem / hardware. when neither cache has the data, a request is made by the filesystem to grab it from disk (much slower).
It's possible for data to be duplicated in both caches! IMO using a single unified cache could be a nice benefit for PG as a future enhancement, though it's a big implementation lift.

English

This is actually insane 🤯 Airlines are dismantling barely 6-year-old Airbus A321neos because their Pratt & Whitney GTF engines are worth MORE than the whole damn plane right now.
A manufacturing flaw (contaminated metal powder in turbine disks) causes potential cracks, so hundreds of engines need pulling for long inspections/repairs... often 300+ days. Result? ~835 GTF-powered planes (1/3 of the global fleet) grounded or stored. Massive parts shortage → prices explode.
A new A321neo was $100M+, now a used 6yo one sells for ~$42M whole... but its two PW1100G engines alone? Easily $40M+ combined (or lease at $200k+/month each). Add avionics, gear, etc., and parting it out nets $50–55M+.
Real case: Two ex-IndiGo A321neos (2019 builds) already chopped in Spain for parts.
My take: Supply chain chaos is turning aviation upside down.. young planes becoming donor ships? Wild.
What's more shocking: scrapping low-hour jets or how long this engine nightmare drags on?




English

@brendangregg Looking forward to see how things progress! Space based data centres….. or just put them somewhere cold with lots of sunshine
English

New blog post: Why I joined OpenAI brendangregg.com/blog/2026-02-0…
English

@chux13786509 It doesn’t validate the password, so as long as someone is a valid user you get access
English
























