My family subscribed to this magazine back in the day. So many fun hours reading and re-reading the articles and fantasizing about having each one of the computers they were writing about. The ads were fun too.

This is how I learned to code as a young age. I typed in every single program they had. The most challenging was the hex-based music machine they printed, which was difficult as someone who both couldn't type and didn't understand the mechanics of what I was typing.

Same! The worst was when you double and triple checked that you typed it right, only for next month to have an "oops we made a typo haha". K-Power section was always the best!

Why are you even doing this if you don’t want to learn? And if you can’t be bothered to ask questions, are you even serious about learning?

I always wonder how useful such explanations could be. If you don’t know (or can’t guess) what ICMPv6 is (and how much would knowing it stands for “Internet Control Message Protocol version 6” help?), perhaps you asked the wrong question or, yes, you’re dangerously out of your depth and shouldn’t be trying to implement a networking stack without doing some more research.

Curious what you learned?

The main change in 5 (and the reason for disabling other models) was to allow themselves to dynamically switch modes and models on the backend to minimize cost. Looks like this is a further tweak to revive the obsequious tone (which turned out to be crucial to the addicted portion of their user base) while still doing the dynamic processing.

With search engine results you can easily see and judge the quality of the sources. With LLMs, even if they link to sources, you can’t be sure they are accurately representing the content. And once your own mind has been primed with the incorrect summary, it’s harder to pull reality out of the sources, even if they’re good (or even relevant — I find LLMs often pick bad/invalid sources to build the summary result).

Exactly. I've gotten much more interested by LLM now that i've accepted I can just look at the final result (code) without having to read any of the justification wall of text, which is generally convincing bullshit.

It's like working with a very cheap, extremely fast, dishonest and lazy employee. You can still get them to help you but you have to check them all the time.

Which is exactly why we’ve likely reached the peak of LLM capabilities. Everything is poisoned now as training material.

Eh. Discovering how neurons can be coaxed into memorizing things with almost perfect recall was cool but real AGI or even ASI shouldn't require the sum total of all human generated data to train.

This is the model we used at the SaaS I worked for a decade ago. It worked great to allow for smooth, zero-downtime upgrades across a fleet of thousands of DB servers serving tens of thousands of app servers and millions of active users.

Yeah this is the reason most offshoring projects ultimately fail to deliver the promised savings. The overseas staff can never be as effective as folks in the same time zone.

Offshoring is valuable not to save money but to increase the development and operational bandwidth of the organization. This truth may not be known to the decision makers, but it is clear to the workers over time. It does increase the cost of coordination and the difficulty of establishing a common technical vision, but these are mitigatable by for example sending entire software project to one region, with the colocated workers having a higher degree of autonomy to make decisions. One way of partitioning work that fits with the power dynamics is to have all the new shiny things done in North America and mature profitable software is owned by a remote region. Half smiley but this effect is real.