HackerTrans
トップ新着トレンドコメント過去質問紹介求人

throwaw33333434

no profile record

コメント

throwaw33333434
·2 年前·議論
my 2016 gate use rolling codes.
throwaw33333434
·2 年前·議論
It breaks when you do a down round. (A for 50mln, exit for 10)
throwaw33333434
·2 年前·議論
Why? It's hard to argue without the details, but I think the direction is good.

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.
throwaw33333434
·2 年前·議論
Go compilator is fast, because it doesn't allow circular references.
throwaw33333434
·3 年前·議論
I know a few people that started a startup, reach a growth limit and it's just a regular business. They work 15h/week max.
throwaw33333434
·3 年前·議論
I'm building a new house and I'm based in Poland. I'm going with air heat-pump* + a simplest fireplace as a backup - not connected to central heating, no electronics. On top of that 10-20kW of solar and a place for batteries. I'll install them once they get cheaper. The house is 200m^2, so I'm hoping to pull only 200 kW/month from the grid and generate rest on my own.

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.
throwaw33333434
·3 年前·議論
This is such a shame, they were planning to built SMRs in my country (Poland). I hope GE Hitachi won't abandon the project now.

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.
throwaw33333434
·3 年前·議論
moloch in ginsberg's sense
throwaw33333434
·3 年前·議論
I bought a new MB CLA 6 years ago, done almost 100k. Nothing broke, never had a problem with software or hardware. Only maintenance was oil, brake fluid, break pad, filters, summer tyres, windshield wipers. Batter is still good, next year I'll replace break discs.

I've been tracking, drifting and driving quite hard (bloody potholes!). I don't think I'll ever sell this car.
throwaw33333434
·3 年前·議論
I like how in German lists US car companies are near bottom, but in US lists it's the German brands.
throwaw33333434
·3 年前·議論
Greek Catholic and for the first time Ukraine.
throwaw33333434
·3 年前·議論
after reading the title I thought the manufacturer coated the adapter in chilly, so cats won't chew on it and thought - "That's a brilliant idea".
throwaw33333434
·3 年前·議論
META: is M3 Pro good enough to run Cyberpunk 2077 smoothly? Does Max really makes a difference?
throwaw33333434
·3 年前·議論
If I understand correctly apart from hardcoded `ifs` there was a backdoor as well.

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

In my book you could argue a criminal case.
throwaw33333434
·3 年前·議論
The other issue is that potholes my get quite expensive.
throwaw33333434
·3 年前·議論
The tragedy of Iran Air Flight 655, which was shot down by the USS Vincennes in 1988, highlights significant user interface (UI) flaws in the Aegis Combat System used by the U.S. Navy. Here's a breakdown of the key UI problems that contributed to this incident:

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.
throwaw33333434
·3 年前·議論
The learn more about this subject google "Argentinian pegging"
throwaw33333434
·3 年前·議論
How did it make it to read pdfs? I tried to create a tool that would extract tables from pdf and covert into CSV. https://chat.openai.com/g/g-e3jhasATL-pdf-ninja but it fails.

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} .... """
throwaw33333434
·3 年前·議論
is it any good for pdfs that are NOT images?
throwaw33333434
·3 年前·議論
Anyone has a way to improve pdf data extraction? I want to covert a table in pdf to a CSV.

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?