I did similar testing with ball-point pens. Eventually I ran into the ISO 12757-2 standard on archival ink, thinking that it would be a great idea to use such pens for my intricate drawings.
Turns out the drawings, some of which I actually sold, faded into oblivion within about a year. After slightly more careful reading of the actual standard, I learned that the drawings were supposed to be archived, i.e. kept in a box or a drawer, and not to be framed for full-time viewing pleasure.
The typical blue ink in the famous BIC ball-point pens (i.e. non-ISO 12757) turns black after some time of sunlight exposure, which seems fine.
I know very little about ink, but I'm a little surprised, especially for something marked as archival.
My only point of comparison is an etching which is a few hundred years old. The ink has no obvious signs of fading. It had some light restoration about forty years ago but we usually have it on display, at least in my life time.
When I've seen other prints of the same etching in galleries, they are usually in the same or slightly worse condition despite, probably, being kept much more carefully.
Thanks for mentioning this. I'm just getting into ballpoints in my art (literally just posted my latest piece online this morning, using Schneider Slider pens which use ink conforming to ISO 12757-2) and have diligently sourced 'archival' inks without looking deeply enough into what it means (a lot of online discussions imply that it's the same as lightfastness).
About BIC, I thought I'd seens some tests online showing that over time the ink yellowed and eventually disappeared, so I've been avoiding them.
Clearly I need to look more deeply into it, especially for work I might sell.
With ball-point ink, I think (but haven't checked recently) that there is unfortunately no safe option.
Some contemporary artists sell digital prints, but if you are selling the real deal, at least ensure that it does not sit in full sunlight all day, because it will fade quickly.
In most museums original prints and drawings are often exhibited for short periods in darkened rooms. I think this takes away from the pleasure, and I prefer to see my art as perishables, which one can enjoy for a few years. An additional problem is that paper is terribly fragile stuff. Framing it (acid free paper) properly (with acid free tape) helps, but can be very expensive and will affect the aesthetics of course.
Some collectors of printmaking art store everything in a drawer and take it out once a year to enjoy it with a good glass of wine.
I really enjoyed reading this even though I have no direct interest in the lightfastness of the pigments in colored pencils.
It was just fun to see what someone who is deeply invested thought important to test, explain and research about something I'd have previously called a matter of aesthetic preference (as opposed to a thing you can benchmark).
There's more high quality engineering discipline in this 'non-engineering' article than in seemingly a lot of self-professed software engineering today ;)
I'm not surprised. As $DAYJOB involves a reasonable amount of requirements for colour accurate previewing in a print context I still feel like I'm never as sure about all steps of the colour pipeline as I should be, and this adds a whole new area to know I'll never feel like I know enough about.
Sadly, I'm not surprised (looks like the site was written twenty years ago). It must be one of those "roulette" ones, because I have not encountered it (or Safari kills it), but it does have that stupid region popup.
As someone who is heavily tattooed, I'd LOVE to see this analysis for tattoo inks.
Fun fact: UV light makes tattoo particles smaller, which makes them easier for your lymphatic system to carry them to your lymph nodes. The particles are easy to transport into the lymph nodes, but difficult for your body to remove from your lymph nodes, meaning that for heavily tattooed people like myself, surgeries can be a potentially very colorful endeavor! (Or, if you have primarily black tattoos, it can be a spooky endeavor, I suppose.)
What a effort.
I did something similar for some pens and different inkjet colors a long time ago, but not nearly as broad or as methodical.
The inkjet inks (especially red) were already blown after a short time (>4 months).
But black still holds up 20 years later till now, only a little bit faded, so one can see the tracks of each row.
Foils (laminate or adhesive foils) or protective spray (UV) did not change the result at all.
But one film tore and gave the whole thing an interesting, crackle-like appearance.
However, the colors all faded in the same way, whatever protective used compared to direct exposure to sunlight.
I've noticed this as well -- at one point I had switched from Canon inkjet refills to a generic third party refills, and years later the photos from the knockoff have faded to an astonishing degree while the Canon ink prints look bright and vibrant.
Going through old photo albums that my parents have show a lot of fading as well, even though the pictures themselves were kept in photo albums in the dark for many years. We have negatives for some of those photos, which when scanned are bright and vibrant, but the prints vary significantly.
uh-oh, did you just anecdotally confirm that buying first party inks is the better decision? careful, you might just confirm to the makers that their DRM policies are legit.
I always say the same thing. HP and Xerox run their (ink and toner) labs for a reason, and their ink might be expensive but it's high quality and dependable.
I seldom print photos with my entry level HP inkjet, using HP inks. Even though my printer uses dye based color cartridges, none of the photos I print have faded even though they are framed and displayed 7/24/365. Ink is not colored water.
If you print/read/shred in two days, 3rd party ink and lower quality paper is OK, but if you want to be able to read or look at that thing after 10+ years, you need better quality paper and inks.
I use fountain pens a lot, and difference between ink quality becomes much more evident there. There are writing inks, there are archival inks (which are not Indian inks), and they behave totally different.
Third party ink is crap. On various forums you will read stories about people who attempt borderless printing and have terrible inksplosions. Quite often they are using third party inks and get no sympathy from people who are serious about inkjet printing.
Inkjet manufacturers put a lot of effort into ink formulations and often these are better in terms of VoC and other parameters. Sometimes you find certain first party inks are not at all lightfast (like the ink for the Epson ET-3750 which I found fades badly in less than six months) but there is very little independent testing of ink performance. If there was there might be a market for ink that performs better than first party ink.
It's a bit of a pity because the printer makers have been trying so many scams that their actual valid selling points are also assumed to be a bamboozlement.
But also, it's a fairly niche requirement to have a non-photo printed document need survive sunlight for years. I'd guess 95% plus of pages that get printed on a normal office printer have minimal need for colour precision, use crappy paper and go in the bin within the year or end up in an archive box, likely never to be seen by human eyes again.
If I remember correctly, HP produces pigment based color inks for their more serious inkjet lines, and these systems generally use 4 or more cartridges.
However, even though I balk at my 4515 for having dye based color inks, their fade resistance is astonishing.
HP always uses pigment based black inks on every inkjet, so text is always at archival quality on good paper.
I still like the tank so I upgraded to an ET-8550 which uses mainly dye-based inks except for a pigment black. Photos and art reproductions I print from that still look OK 2.5 years later of being on display, even in a south-facing room in my house which seems to be a good "accelerated aging" environment. Epson makes an Pro ET line which uses pigment-based ink, but when I tried to get one during the pandemic I couldn't find one so I got the ET-8550.
I did this for a single color from a single printer—the black toner from my Brother laser printer. I left it in my West facing office window for about 18 months. On the BACK SIDE I labeled it with pen. The pen on the back faded to almost nothing but the toner did not fade at all.
I did not do monthly scans, that would have been a better "experiment", but I was satisfied that a B&W laser print would last a very long time.
Maybe I should lightfast test my Brother Laser and my HP Inkjet (with Black Pigment based ink).
I thought that pigment based inks would be both waterproof and lightfast. Since I started to airbrush watercolor over my HP prints I am now very aware that these pigment based inks are not waterproof, even after long drying times.
There's a book where the author tested watercolors which my watercolor teacher said was extremely controversial when it came out, but she was able to replicate some of his results, and so changed which colors she bought.
I tried the colors she recommended, and got good results. I also tested various black pens, and found (25 years ago!) that Micron pens were colorfast. Some black (gel, IIRC) pens faded to a nice sepia.
Pinks and oranges are notoriously not lightfast. Unless you use cadmium, which is kind of not something you want to be handing out to people who might be licking their pencils. Ever notice why yellow cars always end up looking pale lemon-yellow? Lightfast non-toxic cheap red pigments aren't that easy.
Many rose/fuschia colors are not lightfast. Pigments that are not permanent are called 'fugitive' in the arts. Rose madder and alizarin crimson are not lightfast. So are a lot of others: cochineal, geranium lake. That's why you see 'green' cherubs when you visit the Getty museum also why Van Gogh's paintings are not the color they were when he painted them. Most reds for art are now quinacridone or cadmium.
One of Windsor and Newton's (oldschool paint supply manufacturer, its fun to browse history on their website) most expensive paints is 'rose madder geniune.' They claim it is permanent but I don't know how they did it.
It's interesting how pink tends to be the worst of the colors according to the charts. I wonder what it really looks like on paper. Does it disappear completely so it is very hard to see, or is it just invisible in the data as it rounds to 0 but leaves something visible on the paper? If you did an image in pink duotone with the worst offenders, would you have a blank sheet of paper after 26 weeks? Or does it look like something done in white duotone?
Look at all the red-white-and-blue bumper stickers. They are usually white-and-blue.
Around here, we have school buses with a sticker on the back, announcing that they don't turn right on RED (with "RED" being in heavy letters, and colored red).
They frequently say that they don't turn right on.
Red ink absorbs high energy photons so yeah it tends to get trashed. Mineral reds are good (the red on a red metal barn) and some chemical reds but not others. I remember on Jan 1, 2002 walking around Binghamton where there were many white-grey-and-blue flag lawn signs all over town that had been hurriedly installed the prior September and going in to watch a parade on TV where an announcer said “You know what my favorite bumper sticker is? These colors don’t run”
The flag I got from the US Army in 1993 when my father passed away is still in great shape while many of them don’t look so good.
You can (could?) see it in a lot of red cars too, they fade to a colour that I can only describe as "faded red car". They either use a different, more stable paint nowadays or avoid red paint in general.
It's my understanding that natural pigments are in general less colorfast than synthetic ones. If one prefers the color quality of natural pigments, would choosing the "worst" pencils here be indicative of that?
3 images explaining the importance of lightfastness
1 image explaining the measurement process
1 image linking to another article diving much deeper into the methodology
1 image linking to another article on a different color pencil concern (layering)
1 image representing each brand-line's lightfastness
Every single one of those images seemed relevant to the concept presented and clarified something that would have been difficult to articulate succinctly in writing. For example, the "how was this measured" is a lot easier to understand once you've seen the grid of squares before and after than it would be to try and articulate the fading of colors in small squares in text.
There's LOTS of individual images on specific brands, but given their wild degree of variance, I think it's really useful to perceptually see what's going on with each one.
I'm curious, where do you feel the images were "spammy"? It's a conclusion I heartily disagree with, but would love to understand.
I think gennarro is reacting to the very SEO-friendly organization of the article. Every content farm produces articles with this kind of flow, often with Wikipedia-style tables of contents at the beginning. But they do it because it’s very similar to the structure an actually informative article would take! So we can’t tell for sure whether the author adopted an SEO-friendly structure for her informative and original content, or if her content just happens to be a good model of the style that content farms have chosen to imitate.
Accelerated testing (6hrs/day) is standard practice in materials science - it compresses years of normal exposure into months while maintaining relative degradation patterns. Fixatives might alter results by adding UV inhibitors, but most artists want to know worst-case baseline performance.
Although someone will challenge me on that, I'm 100% sure that large chunks of the text are AI-generated. That said, the website itself has been around at least since 2017 (the text just wasn't as verbose - e.g., https://sarahrenaeclark.com/diy-gift-bag/).
So, I suspect it's legit. It's a case of an author leaning on a crutch for writing, but we're here to judge the results, not the phasing.
First, it just reads that way. It's the default style if you ask ChatGPT to write a couple of paragraphs that explain why lightfastness is important.
Second, while I know there are reasons to be skeptical about AI text checkers, the author's earlier (less verbose) style doesn't get flagged at all, while the style in more recent articles gets classified as heavily AI-assisted.
> First, it just reads that way. It's the default style if you ask ChatGPT to write a couple of paragraphs that explain why lightfastness is important.
It doesn't read that way to me, and I've read lots of ChatGPT text. We've come to opposite conclusions, I'm curious what qualities you are identifying/keying off of?
In our studies of ChatGPT's grammatical style (https://arxiv.org/abs/2410.16107), it really loves past and present participial phrases (2-5x more usage than humans). I didn't see any here in a glance through the lightfastness section, though I didn't try running the whole article through spaCy to check. In any case it doesn't trip my mental ChatGPT detector either; it reads more like classic SEO writing you'd see all over blogs in the 20-teens.
edit: yeah, ran it through our style feature tagger and nothing jumps out. Low rate of nominalizations (ChatGPT loves those), only a few present participles, "that" as subject at a usual rate, usual number of adverbs, etc. (See table 3 of the paper.) No contractions, which is unusual for normal human writing but common when assuming a more formal tone. I think the author has just affected a particular style, perhaps deliberately.
Tangent, but I'm curious about how your style feature tagger got "no contractions" when the article is full of them. Just in the first couple of paras we have it's, that's, I've, I'd...
Probably because the article uses the Unicode right single quotation mark instead of apostrophes, due to some automated smart-quote machinery. I'll have to adjust the tagger to handle those.
I did similar testing with ball-point pens. Eventually I ran into the ISO 12757-2 standard on archival ink, thinking that it would be a great idea to use such pens for my intricate drawings.
Turns out the drawings, some of which I actually sold, faded into oblivion within about a year. After slightly more careful reading of the actual standard, I learned that the drawings were supposed to be archived, i.e. kept in a box or a drawer, and not to be framed for full-time viewing pleasure.
The typical blue ink in the famous BIC ball-point pens (i.e. non-ISO 12757) turns black after some time of sunlight exposure, which seems fine.
I know very little about ink, but I'm a little surprised, especially for something marked as archival.
My only point of comparison is an etching which is a few hundred years old. The ink has no obvious signs of fading. It had some light restoration about forty years ago but we usually have it on display, at least in my life time.
When I've seen other prints of the same etching in galleries, they are usually in the same or slightly worse condition despite, probably, being kept much more carefully.
Thanks for mentioning this. I'm just getting into ballpoints in my art (literally just posted my latest piece online this morning, using Schneider Slider pens which use ink conforming to ISO 12757-2) and have diligently sourced 'archival' inks without looking deeply enough into what it means (a lot of online discussions imply that it's the same as lightfastness).
About BIC, I thought I'd seens some tests online showing that over time the ink yellowed and eventually disappeared, so I've been avoiding them.
Clearly I need to look more deeply into it, especially for work I might sell.
With ball-point ink, I think (but haven't checked recently) that there is unfortunately no safe option.
Some contemporary artists sell digital prints, but if you are selling the real deal, at least ensure that it does not sit in full sunlight all day, because it will fade quickly.
In most museums original prints and drawings are often exhibited for short periods in darkened rooms. I think this takes away from the pleasure, and I prefer to see my art as perishables, which one can enjoy for a few years. An additional problem is that paper is terribly fragile stuff. Framing it (acid free paper) properly (with acid free tape) helps, but can be very expensive and will affect the aesthetics of course.
Some collectors of printmaking art store everything in a drawer and take it out once a year to enjoy it with a good glass of wine.
I really enjoyed reading this even though I have no direct interest in the lightfastness of the pigments in colored pencils.
It was just fun to see what someone who is deeply invested thought important to test, explain and research about something I'd have previously called a matter of aesthetic preference (as opposed to a thing you can benchmark).
> test, explain and research about something
There's more high quality engineering discipline in this 'non-engineering' article than in seemingly a lot of self-professed software engineering today ;)
Heh. As a [former][0] artist (and musician), and an engineer, I can confirm that the Venn diagrams overlap, quite a bit.
Pigment color is a real heavy-duty field. There's a guy named Michael Wilcox[1] that is famous for his work on pigment color.
[0] https://news.ycombinator.com/item?id=40917886
[1] https://michaelwilcoxschoolofcolour.com/about-michael-wilcox... (Has an annoying popup on every page).
Thank you for these links!
> Pigment color is a real heavy-duty field
I'm not surprised. As $DAYJOB involves a reasonable amount of requirements for colour accurate previewing in a print context I still feel like I'm never as sure about all steps of the colour pipeline as I should be, and this adds a whole new area to know I'll never feel like I know enough about.
Michael Wilcox's site has virus on every click or link it redirect to a casino or some other site at the end.
Sadly, I'm not surprised (looks like the site was written twenty years ago). It must be one of those "roulette" ones, because I have not encountered it (or Safari kills it), but it does have that stupid region popup.
Here's a video interview with him: https://www.youtube.com/watch?v=nd3hCBPqzYU
This is a really nice work product. I totally appreciate the value of this.
As someone who is heavily tattooed, I'd LOVE to see this analysis for tattoo inks.
Fun fact: UV light makes tattoo particles smaller, which makes them easier for your lymphatic system to carry them to your lymph nodes. The particles are easy to transport into the lymph nodes, but difficult for your body to remove from your lymph nodes, meaning that for heavily tattooed people like myself, surgeries can be a potentially very colorful endeavor! (Or, if you have primarily black tattoos, it can be a spooky endeavor, I suppose.)
What a effort. I did something similar for some pens and different inkjet colors a long time ago, but not nearly as broad or as methodical. The inkjet inks (especially red) were already blown after a short time (>4 months). But black still holds up 20 years later till now, only a little bit faded, so one can see the tracks of each row.
Foils (laminate or adhesive foils) or protective spray (UV) did not change the result at all. But one film tore and gave the whole thing an interesting, crackle-like appearance. However, the colors all faded in the same way, whatever protective used compared to direct exposure to sunlight.
I've noticed this as well -- at one point I had switched from Canon inkjet refills to a generic third party refills, and years later the photos from the knockoff have faded to an astonishing degree while the Canon ink prints look bright and vibrant.
Going through old photo albums that my parents have show a lot of fading as well, even though the pictures themselves were kept in photo albums in the dark for many years. We have negatives for some of those photos, which when scanned are bright and vibrant, but the prints vary significantly.
uh-oh, did you just anecdotally confirm that buying first party inks is the better decision? careful, you might just confirm to the makers that their DRM policies are legit.
I always say the same thing. HP and Xerox run their (ink and toner) labs for a reason, and their ink might be expensive but it's high quality and dependable.
I seldom print photos with my entry level HP inkjet, using HP inks. Even though my printer uses dye based color cartridges, none of the photos I print have faded even though they are framed and displayed 7/24/365. Ink is not colored water.
If you print/read/shred in two days, 3rd party ink and lower quality paper is OK, but if you want to be able to read or look at that thing after 10+ years, you need better quality paper and inks.
I use fountain pens a lot, and difference between ink quality becomes much more evident there. There are writing inks, there are archival inks (which are not Indian inks), and they behave totally different.
Third party ink is crap. On various forums you will read stories about people who attempt borderless printing and have terrible inksplosions. Quite often they are using third party inks and get no sympathy from people who are serious about inkjet printing.
Inkjet manufacturers put a lot of effort into ink formulations and often these are better in terms of VoC and other parameters. Sometimes you find certain first party inks are not at all lightfast (like the ink for the Epson ET-3750 which I found fades badly in less than six months) but there is very little independent testing of ink performance. If there was there might be a market for ink that performs better than first party ink.
It's a bit of a pity because the printer makers have been trying so many scams that their actual valid selling points are also assumed to be a bamboozlement.
But also, it's a fairly niche requirement to have a non-photo printed document need survive sunlight for years. I'd guess 95% plus of pages that get printed on a normal office printer have minimal need for colour precision, use crappy paper and go in the bin within the year or end up in an archive box, likely never to be seen by human eyes again.
If I remember correctly, HP produces pigment based color inks for their more serious inkjet lines, and these systems generally use 4 or more cartridges.
However, even though I balk at my 4515 for having dye based color inks, their fade resistance is astonishing.
HP always uses pigment based black inks on every inkjet, so text is always at archival quality on good paper.
I still like the tank so I upgraded to an ET-8550 which uses mainly dye-based inks except for a pigment black. Photos and art reproductions I print from that still look OK 2.5 years later of being on display, even in a south-facing room in my house which seems to be a good "accelerated aging" environment. Epson makes an Pro ET line which uses pigment-based ink, but when I tried to get one during the pandemic I couldn't find one so I got the ET-8550.
Crazy amount of effort here. Awesome.
I did this for a single color from a single printer—the black toner from my Brother laser printer. I left it in my West facing office window for about 18 months. On the BACK SIDE I labeled it with pen. The pen on the back faded to almost nothing but the toner did not fade at all.
I did not do monthly scans, that would have been a better "experiment", but I was satisfied that a B&W laser print would last a very long time.
Maybe I should lightfast test my Brother Laser and my HP Inkjet (with Black Pigment based ink).
I thought that pigment based inks would be both waterproof and lightfast. Since I started to airbrush watercolor over my HP prints I am now very aware that these pigment based inks are not waterproof, even after long drying times.
There's a book where the author tested watercolors which my watercolor teacher said was extremely controversial when it came out, but she was able to replicate some of his results, and so changed which colors she bought.
I tried the colors she recommended, and got good results. I also tested various black pens, and found (25 years ago!) that Micron pens were colorfast. Some black (gel, IIRC) pens faded to a nice sepia.
And OH YEAH THE BOOK: Michael Wilcox's "The Wilcox Guide to the Best Watercolor Paints".
I did this for colored mechanical pencil leads. Short answer, use the Staedtler Mars Micro Color if you need lightfastness.
Pinks and oranges are notoriously not lightfast. Unless you use cadmium, which is kind of not something you want to be handing out to people who might be licking their pencils. Ever notice why yellow cars always end up looking pale lemon-yellow? Lightfast non-toxic cheap red pigments aren't that easy.
Many rose/fuschia colors are not lightfast. Pigments that are not permanent are called 'fugitive' in the arts. Rose madder and alizarin crimson are not lightfast. So are a lot of others: cochineal, geranium lake. That's why you see 'green' cherubs when you visit the Getty museum also why Van Gogh's paintings are not the color they were when he painted them. Most reds for art are now quinacridone or cadmium.
One of Windsor and Newton's (oldschool paint supply manufacturer, its fun to browse history on their website) most expensive paints is 'rose madder geniune.' They claim it is permanent but I don't know how they did it.
Note: x 'lake' means x dye turned into a pigment.
It's interesting how pink tends to be the worst of the colors according to the charts. I wonder what it really looks like on paper. Does it disappear completely so it is very hard to see, or is it just invisible in the data as it rounds to 0 but leaves something visible on the paper? If you did an image in pink duotone with the worst offenders, would you have a blank sheet of paper after 26 weeks? Or does it look like something done in white duotone?
Red, in general, sucks.
Look at all the red-white-and-blue bumper stickers. They are usually white-and-blue.
Around here, we have school buses with a sticker on the back, announcing that they don't turn right on RED (with "RED" being in heavy letters, and colored red).
They frequently say that they don't turn right on.
Red ink absorbs high energy photons so yeah it tends to get trashed. Mineral reds are good (the red on a red metal barn) and some chemical reds but not others. I remember on Jan 1, 2002 walking around Binghamton where there were many white-grey-and-blue flag lawn signs all over town that had been hurriedly installed the prior September and going in to watch a parade on TV where an announcer said “You know what my favorite bumper sticker is? These colors don’t run”
The flag I got from the US Army in 1993 when my father passed away is still in great shape while many of them don’t look so good.
You can (could?) see it in a lot of red cars too, they fade to a colour that I can only describe as "faded red car". They either use a different, more stable paint nowadays or avoid red paint in general.
It's my understanding that natural pigments are in general less colorfast than synthetic ones. If one prefers the color quality of natural pigments, would choosing the "worst" pencils here be indicative of that?
In short, Caran d'Ache has a name and that price tag for a reason.
Sad. I was raised that Prismacolors were the end all be all of colored pencils.
https://web.archive.org/web/https://sarahrenaeclark.com/ligh...
Project Farm: art edition. I love it!
Very spammy images. Is this a legit resource? The test is interesting though 6 hrs/day in sunlight is very extreme for artwork.
Also I’m wondering is a fixer would help or hurt the testing. This is common with some art, like pastels.
There's:
1 header image
1 image showing in process
1 image explaining lightfastness
3 images explaining the importance of lightfastness
1 image explaining the measurement process
1 image linking to another article diving much deeper into the methodology
1 image linking to another article on a different color pencil concern (layering)
1 image representing each brand-line's lightfastness
Every single one of those images seemed relevant to the concept presented and clarified something that would have been difficult to articulate succinctly in writing. For example, the "how was this measured" is a lot easier to understand once you've seen the grid of squares before and after than it would be to try and articulate the fading of colors in small squares in text.
There's LOTS of individual images on specific brands, but given their wild degree of variance, I think it's really useful to perceptually see what's going on with each one.
I'm curious, where do you feel the images were "spammy"? It's a conclusion I heartily disagree with, but would love to understand.
I think gennarro is reacting to the very SEO-friendly organization of the article. Every content farm produces articles with this kind of flow, often with Wikipedia-style tables of contents at the beginning. But they do it because it’s very similar to the structure an actually informative article would take! So we can’t tell for sure whether the author adopted an SEO-friendly structure for her informative and original content, or if her content just happens to be a good model of the style that content farms have chosen to imitate.
Accelerated testing (6hrs/day) is standard practice in materials science - it compresses years of normal exposure into months while maintaining relative degradation patterns. Fixatives might alter results by adding UV inhibitors, but most artists want to know worst-case baseline performance.
She didn't even promoted a single pencil brand on the conclusion. She just shows the data and let the viewers decide.
What did you find spammy about the images? The ads for the artist's coloring books and calendars and such?
Although someone will challenge me on that, I'm 100% sure that large chunks of the text are AI-generated. That said, the website itself has been around at least since 2017 (the text just wasn't as verbose - e.g., https://sarahrenaeclark.com/diy-gift-bag/).
So, I suspect it's legit. It's a case of an author leaning on a crutch for writing, but we're here to judge the results, not the phasing.
Why do you think it's AI? It reads to me like someone who has a special interest and a data driven mindset.
I've seen plenty of people "rate every X" in youtube videos or blogs before, this one is just more data oriented than most.
First, it just reads that way. It's the default style if you ask ChatGPT to write a couple of paragraphs that explain why lightfastness is important.
Second, while I know there are reasons to be skeptical about AI text checkers, the author's earlier (less verbose) style doesn't get flagged at all, while the style in more recent articles gets classified as heavily AI-assisted.
> First, it just reads that way. It's the default style if you ask ChatGPT to write a couple of paragraphs that explain why lightfastness is important.
It doesn't read that way to me, and I've read lots of ChatGPT text. We've come to opposite conclusions, I'm curious what qualities you are identifying/keying off of?
In our studies of ChatGPT's grammatical style (https://arxiv.org/abs/2410.16107), it really loves past and present participial phrases (2-5x more usage than humans). I didn't see any here in a glance through the lightfastness section, though I didn't try running the whole article through spaCy to check. In any case it doesn't trip my mental ChatGPT detector either; it reads more like classic SEO writing you'd see all over blogs in the 20-teens.
edit: yeah, ran it through our style feature tagger and nothing jumps out. Low rate of nominalizations (ChatGPT loves those), only a few present participles, "that" as subject at a usual rate, usual number of adverbs, etc. (See table 3 of the paper.) No contractions, which is unusual for normal human writing but common when assuming a more formal tone. I think the author has just affected a particular style, perhaps deliberately.
Tangent, but I'm curious about how your style feature tagger got "no contractions" when the article is full of them. Just in the first couple of paras we have it's, that's, I've, I'd...
Probably because the article uses the Unicode right single quotation mark instead of apostrophes, due to some automated smart-quote machinery. I'll have to adjust the tagger to handle those.