4 ms·
Is this a GPT-generated article? I see lots of weird patterns like the following, from the section on HTTP/3: >QUIC uses loss recovery to recover from packet
by nvy 4y ago
Is this a GPT-generated article? I see lots of weird patterns like the following, from the section on HTTP/3:
>QUIC uses loss recovery to recover from packet loss. QUIC uses a combination of congestion control and loss recovery to recover from packet loss. This is similar to TCP where TCP uses a combination of congestion control and loss recovery to recover from packet loss.
- KingMob 4y agoI hadn't thought of that, but maybe. There's some weird inaccuracies. The paragraphs on HTTP/1.0 makes it sound like images weren't sent before HTTP/1.0, but NCSA Mosaic was doing exactly that before HTTP/1.0. The whole bit about HTTP/2 server push neglects to mention that it's been effectively deprecated, which seems kind of like an important point. Likewise, for request prioritization, it leaves out that HTTP/2's method is frequently worse than nothing (without a LOT of careful work, measurement, and tuning), and that HTTP/3 vastly simplified it.
- postsantum 4y agoYeah, the entire website looks like a content farm
- endorphine 4y agoJust came here to ask the same thing. It feels like a low-quality article to me - a lot of things are repeated and just feels like it's a copy-paste collection of snippets from various sources without the author understanding what it really talks about.
- 2bluesc 4y agoThere's a lot of repeated stuff for no reason. The congestion control and loss recovery sections are a struggle to read. Ctrl-f "uses a combination of congestion control and loss recovery to recover from packet loss" or "This is similar to"
- chrismorgan 4y agoVarious errors about the place too, e.g. the “HTTP is TCP” thing near the start (twice), and the HTTP/1.0 request sample includes a Host header (that came later), and has a User-Agent value suggesting a 2017 browser.