I want to know more about the products I use. A can of coke should have a QR code with all the information about the product. Which mines provided the Aluminium?

Good programmers log.

Also, I've been wondering how much efficiency would I get if I could put some mirror and direct the light to the unit in the winter. It could help with defrosting as well.

*I was considering horizontal heat pump, but it's not worth it. It affects vegetation and future construction.

Currently, I'm very enthusiastic about fervoenergy.com. It presents an outstanding alternative to gas and coal for district heating, and it may even have potential for electricity generation.

I've been tracking, drifting and driving quite hard (bloody potholes!). I don't think I'll ever sell this car.

Russian agencies could use it to slow down transit of military aid to Ukraine.

In my book you could argue a criminal case.

IFF Console Operation Flaw: The aircraft identification friend or foe (IFF) system, crucial for distinguishing between civilian and military aircraft, had a critical operational flaw. Although Flight 655's IFF correctly identified it as a civilian airliner, the Aegis system required operators to manually follow the contact with a cursor. In this instance, the operator did not "slave" the cursor to Flight 655, causing the system to query the IFF of a different aircraft, an Iranian F-14, from the same location.

Dashboard Design Issue: Aegis' large displays, used for senior decision-making, failed to show critical information like altitude. This omission made it challenging to assess the intent of an approaching aircraft. Flight 655's climbing trajectory, a non-threatening maneuver, wasn't immediately apparent to the captain from these displays.

Tracking Number Confusion: Aegis' feature of unifying data across multiple ships led to a fatal misunderstanding. Both the Vincennes and its escort, USS Sides, tracked Flight 655 but assigned it different tracking numbers. When Aegis unified these contacts, it chose one tracking number and recycled the other, which was then assigned to a U.S. bomber. This led to confusion about the altitude of Flight 655, as the captain, unaware of the tracking number change, received data about the wrong aircraft, believing it was descending rapidly.

These UI issues, combined with the high-stress environment and other factors, contributed to the crew's misunderstanding of the actual situation, ultimately leading to the tragic decision to fire on the civilian airliner. This case underscores the critical importance of intuitive, clear, and comprehensive user interface design, especially in high-stakes environments like military operations.

It works a bit better if I extract the string in python and do some clean up before sending

import fitz # PyMuPDF

pdf_document = fitz.open("foo.pdf")

page_number = 1

page = pdf_document.load_page(page_number - 1)

text = page.get_text("text")

response = client.chat.completions.create( model="gpt-3.5-turbo", messages=[ { "role": "system", "content": f""" ..... {text} .... """

so far the best performance has conversation to string

import fitz # PyMuPDF

pdf_document = fitz.open("foo.pdf") page_number = 1 page = pdf_document.load_page(page_number - 1) text = page.get_text("text")

response = client.chat.completions.create( model="gpt-3.5-turbo", messages=[ { "role": "system", "content": f""" ..... {text} .... """

If I try regular ChatGPT it takes 3 minutes to covert the table (I have to press continue). Is there a way to force API to create whole CSV? some sort of retry?