GianFabien

Natural language is full of nuances and implicit cultural and domain specific connotations. LLMs work because they have been trained on massive corpus of materials that encode some (but not all) of those hidden meanings.

With specification driven development, the prompts need to be very detailed in order to produce exactly what is required.

It took a couple of million years of evolution for some humans to attain the competence you describe. A couple of years of engineering is not going to catch up anytime soon.

Copyright laws didn't even pause the Big AI companies from hoovering up all material to train their systems. With their financial and thus legal resources they would be very expensive to sue for patent infringement.

Not very personal when it probably takes two persons to load it into a rack.

Heading should be edited - its not a PC and not a workstation.

Here we go again! 2020 deja vu! Just in time for USA elections.

Early Reuters and Bloomberg terminals had a text-centric aesthetic.

AFAIK on a PC, when booted into text-only mode, you are still using a bit-mapped framebuffer. From my experiments you can mix graphics and text.

It's the nanny state dynamic. Parents are avoiding responsibility for their own children, instead demanding that the schools, teachers and government take on the unpleasant, confronting and often difficult role.

I've tried this approach, but I found that most email providers aggressively filter out all incoming messages that don't come from validated sources. No matter what I tried deliverability of emails remained very spotty. I did configure DKIM, DMARC, SPIF, etc as suggested. But minimal improvements. Eventually ended up going with AWS SES.

All the data centres I worked in had Halon and in recent times its replacements as fire suppression system. Unlike water, these systems suffocate the fire and once vented the equipment is undamaged. Water, computers and high electrical currents are a very bad combination.

After I read the first paragraph, I could guess the rest. It is typical for the legacy enterprise system space.

My diagnosis is that you care too much. The enterprise IT space is kabuki theatre.

So what if it takes 6 months to accomplish what should take 2 weeks? Explain to the customers that their needs are so special and particular that it takes your highly trained and experienced staff 8 months to accomplish the high quality results that you demand. Every line of code is carefully written and rechecked, etc. No AI slop is allowed anywhere near your exacting group. You are delivering Saab fighter jet plane quality software, not some cheap and nasty jeep software from Cuba.

Of course, the real challenge is to say stuff like that with a straight face. If necessary, take improv acting lessons. After all it is all an absurdist theatre production.

... and keep sending out resumes, etc.

No job is worth jeopardising your mental health.

Technology and automation have never gotten in the way of people pursuing and enjoying hobbies. Painters still paint, woodworkers still make bespoke furniture, etc, etc.

Isn't getting a good night's sleep essential for health reasons?

Every "wasted $" ended up in some parasitical grifter's pocket.

Whilst these sorts of analyses are informative, they lack answers to the who profits? question.

There is an interesting tell, they sent the old presidential plane. But then Trump flew onwards from England on the new one. The two planes are sufficiently different that they can't be mistaken for one another.