Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The cost of transparency: Living with schizoaffective disorder in tech

kennethreitz.org

Post reply on HN
Print view View original

101–110 of 130 posts

Re: The cost of transparency: Living with schizoaffective disorder in tech

#101

Post by JohnBooty »

nemomarx wrote: post #84
JohnBooty wrote: post #78

Earlier quoted context omitted.

I don't mind AI polishing any more than I mind somebody using a spellchecker. The article overwhelmingly reads like a (painfully) honest and personal account of what he's going through, it's not AI slop.

It's definitely not slop, but there's something kind of sad about a very personal and emotional account needing to be fed through the corpo speak grammar checker? At least in my opinion. It feels like if a friend wanted to vent to me on discord but didn't trust me enough to do it without editing their post for hours, or something.

There's a very large range of situations (many highly personal) for which I'd consider an AI writing assist to be super valid. They may have dyslexia. English may not be their first language. They may have brilliant ideas but otherwise struggle with writing. They may simply be highly invested in making sure they are expressing themselves clearly. They may be too physically or mentally tired to polish their writing "manually." They may be under unfortunate time constraints. Maybe the guy has kids to take care of, or just needed to finish his blog post by a certain time so he could get a good night's sleep.

The list of possible reasons is so vast and varied that I coalesce this down into essentially "I don't care and I don't need to know. As long as the end result is an honest representation of the author's intent, i.e. is not AI slop"

    It feels like if a friend wanted to vent to me on discord 
    but didn't trust me enough to do it without editing their 
    post for hours, or something.
I'm being pedantic but as a good comparison. A friend DM'ing you is explicitly asking you to spend your time reading the message. Additionally you and your friend have preexisting context and rapport. If they don't express themselves in anal-retentive detail clearly in the DM, you can fill in the gaps with your shared context, a luxury that a person writing for the public doesn't have. (again, pedantic on my part, I know!)

Re: The cost of transparency: Living with schizoaffective disorder in tech

#102

Post by bradstewart »

morleytj wrote: post #45
bradstewart wrote: post #35

Earlier quoted context omitted.

What specifically is the LLM verbiage you're seeing here? That reads like a normal sentence to me.

It's very common to see the particular syntactic structure of restating a point in the following general manner from Claude/ChatGPT in my experience and that of others: "It's not X -- it's Y." or "This isn't just X -- It's actually Y." Usually with an emdash there as well for the separation. As I said it's very plausibly becoming more common among people not using LLM-assisted writing too, just from seeing the stylis…

Interesting, thanks. I've always been a fairly "heavy" (vs other people) user of the emdash after a high school english teacher made us use one in every paper to learn how they worked (along with a colon), and I've been a fan ever sense.

The "it's not ... it's" phrasing though definitely stands out as a bit odd when repeated.

Re: The cost of transparency: Living with schizoaffective disorder in tech

#103

Post by ModernMech »

nemomarx wrote: post #32
morleytj wrote: post #30

This is very unfortunate and I'm sorry to hear that the author has been excluded and is suffering to this extent. On another note though: "This isn't paranoia—it's pattern recognition honed by lived experience." I can't stop seeing the LLM verbiage everywhere I look. I feel like once you recognize the repeated syntax that got RLHF'd into all of these models you never stop seeing it. Maybe everyone is learning those p…

Yeah, I also kind of wish he published the blog in his own words instead of using what seems to be LLM polishing. It seems ironically like the same defensiveness and lack of transparency he wants to avoid?

As someone who is autistic and bipolar, I've been told my words often come across as unnecessarily aggressive even though I mean them in a neutral way, so I sometimes use LLMs to try and mediate my tone when I'm talking to an unknown audience. I suspect the author has been told he is acerbic and might be self-conscious of using his own words.

Re: The cost of transparency: Living with schizoaffective disorder in tech

#104

Post by kennethreitz »

FYI, I updated the post to acknowledge the public criticisms of my previous behavior:

> I need to be clear about something. I'm not claiming innocence in every workplace conflict or community dispute. I've made mistakes, handled situations poorly, and there's been legitimate criticism of my behavior in some cases. The Requests 3 fundraising situation, in particular, was handled badly - I took on commitments I couldn't deliver and didn't communicate well about the problems. I've apologized for this, though I understand the damage was already done. Mental illness doesn't excuse harmful behavior, and I'm not asking it to.

Thanks for the feedback!

Re: The cost of transparency: Living with schizoaffective disorder in tech

#105

Post by ModernMech »

nemomarx wrote: post #84
JohnBooty wrote: post #78

Earlier quoted context omitted.

I don't mind AI polishing any more than I mind somebody using a spellchecker. The article overwhelmingly reads like a (painfully) honest and personal account of what he's going through, it's not AI slop.

It's definitely not slop, but there's something kind of sad about a very personal and emotional account needing to be fed through the corpo speak grammar checker? At least in my opinion. It feels like if a friend wanted to vent to me on discord but didn't trust me enough to do it without editing their post for hours, or something.

> there's something kind of sad about a very personal and emotional account needing to be fed through the corpo speak grammar checker?

There is! But it of course follows the content, which is about how the author's emotions have caused him professional harm. Someone who's been through that would be shy about sharing their raw words.

Re: The cost of transparency: Living with schizoaffective disorder in tech

#106

Post by nemomarx »

ModernMech wrote: post #103
nemomarx wrote: post #32

Earlier quoted context omitted.

Yeah, I also kind of wish he published the blog in his own words instead of using what seems to be LLM polishing. It seems ironically like the same defensiveness and lack of transparency he wants to avoid?

As someone who is autistic and bipolar, I've been told my words often come across as unnecessarily aggressive even though I mean them in a neutral way, so I sometimes use LLMs to try and mediate my tone when I'm talking to an unknown audience. I suspect the author has been told he is acerbic and might be self-conscious of using his own words.

I would definitely believe that, yeah. I just think that reflects something sad about the culture we're in, especially in an essay about struggling with transparency. I would like to get more of a sense of his emotions and writing style, but maybe that would have consequences in a public post and he's more guarded about that.

Re: The cost of transparency: Living with schizoaffective disorder in tech

#107

Post by kennethreitz »

jokoon wrote: post #46
runjake wrote: post #10

The Python community didn’t exclude Kenneth because of his condition, at least directly. They excluded him because of his actions and history of shady, manipulative behaviors. And while workplaces should be accommodating, there’s a point where it gets to be too much. I worked at a place with a person with similar behaviors and it destroyed that company because they refused to acknowledge reality: this was a business,…

This doesn't surprise me. It's central to understand that mental health can alter perception and judgment, even common mental illnesses like depression and anxiety. I have chronic depression, and even after treating it for 15 years, I still realizes after the fact, despite being attentive to how I behave, that I just think differently when I am depressed. It's very common for mentally ill people to think they are bei…

Truth. This manifests as stigma, however, and needs awareness to compensate for.

Re: The cost of transparency: Living with schizoaffective disorder in tech

#108

Post by kennethreitz »

nemomarx wrote: post #23
sfn42 wrote: post #16

Earlier quoted context omitted.

This blog I found seems to support your view https://vorpus.org/blog/why-im-not-collaborating-with-kennet...

> I think a lot of people don't realize how little Reitz actually has to do with Requests development. For many years now, actual maintenance has been done almost exclusively by other volunteers. If you look at the maintainers list on PyPI, you'll see he doesn't have PyPI rights to his own project, because he kept breaking stuff, so the real maintainers insisted on revoking his access. If you clone the Requests git r…

I didn't lose control of my project, I removed myself from it intentionally (and unprompted).

Re: The cost of transparency: Living with schizoaffective disorder in tech

#109

Post by kennethreitz »

4ggr0 wrote: post #33
sfn42 wrote: post #16

Earlier quoted context omitted.

This blog I found seems to support your view https://vorpus.org/blog/why-im-not-collaborating-with-kennet...

There's an "apology"/response from Keith. https://kennethreitz.org/essays/2023-01-an_overdue_apology "i'm sorry, hope we can continue working together in the future" seems like a very mild response to allegedly fumbling 30k, but what do i know... > the asynchronous landscape within Python's ecosystem failed to meet my expectations, leading me to conclude that Requests should retain its synchronous nature. i concluded…

My name is Kenneth, not Keith.

Re: The cost of transparency: Living with schizoaffective disorder in tech

#110

Post by nemomarx »

kennethreitz wrote: post #108
nemomarx wrote: post #23

Earlier quoted context omitted.

> I think a lot of people don't realize how little Reitz actually has to do with Requests development. For many years now, actual maintenance has been done almost exclusively by other volunteers. If you look at the maintainers list on PyPI, you'll see he doesn't have PyPI rights to his own project, because he kept breaking stuff, so the real maintainers insisted on revoking his access. If you clone the Requests git r…

I didn't lose control of my project, I removed myself from it intentionally (and unprompted).

gotcha, and I appreciate the info. that sounds less acrimonious than I was imagining definitely
Post reply on HN