rmykhajliw

Welcome to Ukraine where local people installed around 8GWt of batteries at home mostly because of huge outages after missile and ballistic terroristic attacks.

Batteries installation is not about money , average LiFePo4 battery life is around 10 years, so you will save nothing, unless you have some local generation like solar panels. In this case you could shrink payback time to 5 years instead of 10 and it makes some sense.

But the biggest problems here 2: 1. there a critical point after installing a huge amount of balance batteries the diff between day/night hours would be mitigated. And prices may down 30%. As result economical factor of installing energy storage systems would be very questionable. However cheap energy will increase energy usage by homes and industry and this may/might/probaly/i_hope. boost economy

2. The second factor riskier because when many people are installing energy storages and become independent of power grid, or even disconnects from the grid, there still exists power grid you have to support. Right now the price of grid support included in energy bill for electric company. The grid usage price is stable , it doesn't matter you support 1 person or a 1mln residents city. So the less people are using THE GRID - the higher prices for the rest people. And those high price forces rest people to disconnect from the grid. In general we may/I_hope_not to have situation with plenty energy islands and no more national grid systems

As obvious it's possible to use local Ollama models instead of ChatGPT , it you have large amount of data. But I love using the same approach even for building linking inside internal documentation, I really helps to keep all the changes updated and related.

Small approach to use RAG for better result in similarity for related content. The best thing is the process of getting vectors are don't only once, then you can reuse them indefinitely.

My point is when you have 100 different ways to achieve result - it's always led to inconsistency. And it's very dangerous in large distributed app because resolution of those small differences means you have to support 50+ variation of the same submit button.

Personally I'd prefer to have up to 5-10 option of buttons for all possible cases. And it's not only visual simlicity, and not only easier to support but even help your own customers because they are not have to looking for 100+ different button styles.

I don’t believe the Tailwind approach is a stable, scalable, or reliable solution for large, long-living products.

It works well for fast delivery, prototypes, and teams optimizing for short-term speed, but over time it tends to spread styling decisions across markup, making the system harder to control, reason about, and maintain.

The main issue is games. Drivers and overall performance under Linux for win games run under Wine is pretty low. The good stuff they became runnable but framerate of 12 fps is nothing for my razor laptop with rtx in Cossacs 3.

Now my setup is MacBook for work, and special game laptop Razer Blade especially for games with win11 removed defender/firewall and tons on useless stuff. Personally for me win7 will be enough but I just cannot install it on my laptop due to lack to drivers.

A list of base hack like content grouping, user id, custom user properties and UX metric like page generation time and page load time

Continue discussion about using static files as database instead of default relational MySql/Postgree/Mongo