The practice of rephrasing to get your word count up really irks me.
"Today, wolves, foxes, and snakes inhabit the cave, but it was once a popular spot for human pastoralists and their domesticated animals."
"Rather, they think it was a convenient spot for herders to stop and provide their flocks with shade and water."
"Even though humans didn’t have a permanent presence in the lava tube, the natural structure provided shelter for people and their herds for thousands of years"
They're a gizmodo writer. An attention predator. If they cared about producing genuine values for humanity, they would not be working in attention predator journalism
Therefore don't expect their writing to improve from your criticism. It's designed to take, not give.
I like to stop reading these kinds of authors as soon as I realize I'm experiencing a memetic attention attack. Then I write a snarky comment so others will start to hate them as well.
Generalizing topics and/or adding caveats about how something may not always be true is the easiest way for a LLM to increase the rate of providing factually accurate responses (aside from refusing to answer). Both of those strategies requires being more verbose.
Concise and to the point is not in the best interest of a LLM designed to give decent answers 99% of the time.
I also believe that OpenAI's RLHF process highly biased the model towards producing that sort of specific verbose, padded-out 'chatGPT speak' that we are seeing. The RLHF fine tuning process took outputs from the instruction-tuned model and A/B tested variations with human workers who may or may not have had the same feelings towards writing quality as many of us.
The process resulted in those verbose, interjection-laden responses that we see now, because that type of response was deemed 'better' (thumbs-up'd more) than the shorter, more-direct-but-less-impressive-sounding responses.
I use it for work so much I’d gladly pay way more than $20 just for access to GPT-4. It’s pretty terrible at programming but it still saves me loads of time generating the easy functions.
Anything remotely complex I still do by hand. But holy shit its nice having something do my boilerplate.
Gizmodo and anything gawker media is utter and total junk, pushing Disney media and other hidden agenda. Was irked when I accidentally clicked it, I block it on almost all my systems to avoid feeding the beast.
The practice of rephrasing to get your word count up really irks me.
"Today, wolves, foxes, and snakes inhabit the cave, but it was once a popular spot for human pastoralists and their domesticated animals."
"Rather, they think it was a convenient spot for herders to stop and provide their flocks with shade and water."
"Even though humans didn’t have a permanent presence in the lava tube, the natural structure provided shelter for people and their herds for thousands of years"
I don't know about you all, but rather than commercial commentary I'd prefer if we submitted the papers: https://news.ycombinator.com/item?id=40092285
Gotta have 8 paragraphs so they can squeeze in 20 ads.
Maybe my dns blacklist needs updating, but the ads were so bad I bailed.
ublock origin browser plugin and I saw zero ads
Congratulations
Maybe, it could also just be professional immaturity. It seems more written by a college student than a seasoned journalist.
The quality of news peaked about 20 years ago after search killed monetization. There is no going back.
They're a gizmodo writer. An attention predator. If they cared about producing genuine values for humanity, they would not be working in attention predator journalism
Therefore don't expect their writing to improve from your criticism. It's designed to take, not give.
I like to stop reading these kinds of authors as soon as I realize I'm experiencing a memetic attention attack. Then I write a snarky comment so others will start to hate them as well.
It is fine enough for me as my English skill is bad enough to notice these things.
to not* notice
The verbosity reminds me of ChatGPT generated text
Where did you think ChatGPT got their training data?
Doesn’t matter.
Generalizing topics and/or adding caveats about how something may not always be true is the easiest way for a LLM to increase the rate of providing factually accurate responses (aside from refusing to answer). Both of those strategies requires being more verbose.
Concise and to the point is not in the best interest of a LLM designed to give decent answers 99% of the time.
I also believe that OpenAI's RLHF process highly biased the model towards producing that sort of specific verbose, padded-out 'chatGPT speak' that we are seeing. The RLHF fine tuning process took outputs from the instruction-tuned model and A/B tested variations with human workers who may or may not have had the same feelings towards writing quality as many of us.
The process resulted in those verbose, interjection-laden responses that we see now, because that type of response was deemed 'better' (thumbs-up'd more) than the shorter, more-direct-but-less-impressive-sounding responses.
My chats usually start with:
>Hi again :wave:.
>As always I prefer terse replies.
>Let's <context of activity>
><First question>
I usually then get short answers and then I query for more info if required.
Set a custom instruction (I think only a ChatGPT paid option?). But $20 a month or whatever is easily worth it for the utility it provides.
I use it for work so much I’d gladly pay way more than $20 just for access to GPT-4. It’s pretty terrible at programming but it still saves me loads of time generating the easy functions.
Anything remotely complex I still do by hand. But holy shit its nice having something do my boilerplate.
Gizmodo and anything gawker media is utter and total junk, pushing Disney media and other hidden agenda. Was irked when I accidentally clicked it, I block it on almost all my systems to avoid feeding the beast.
[dead]
Spoiler alert: it was a shelter for people moving herds between oases, and doesn’t seem to have had a permanent population.
Source paper: https://journals.plos.org/plosone/article?id=10.1371/journal...
The location: https://www.google.com/maps/@25.5879436,39.7629388,1317m/dat...
If you zoom out there are a crapload of lava flows and volcanic craters nearby. A bit to the east there is this landscape: https://www.google.com/maps/@25.6327898,39.9400629,22944m/da...
Reminds me of the Sierra Pinacate on the Sonora/Arizona border. It has human history going back more than 20,000 years.
https://en.m.wikipedia.org/wiki/El_Pinacate_y_Gran_Desierto_...
“Humans sheltered for thousands of years” isn’t quite right. Humans stopped by and used this shelter from time to time for thousands of years.
"I've lived in my house for 20 years"
Depepends, if you want to introduce ambiguitity or remove it.
I went to Cueva de los Verdes in Lanzarote (Canaries) somewhat recently, it's 3.7 miles long (plus another mile under sea level).
https://en.wikipedia.org/wiki/Cueva_de_los_Verdes
Seems like it would make a good sietch.
Looks like Jabbas palace.
And deeper in the cave is worse, including some gruesome human remains
https://gizmodo.com/hyenas-left-a-massive-pile-of-bones-in-a...