Ben Grosser
@bengrosser.bsky.social
Artist, recomposing social platforms, AI futures, and interface cultures. Known for works like Demetricator, Computers Watching Movies, and minus.social. Faculty Associate @ Berkman Klein Harvard, Prof @ UIUC. bengrosser.com
Pinned
Ben Grosser
@bengrosser.bsky.social
· Aug 25
New paper w/ @spold.bsky.social. We introduce new methods to critique linguistic interfaces of AI chatbots, theorize one of ChatGPT's central engagement mechanics as the "praise/prompt envelope," and preview artworks that attune users to see its aesthetics of engagement dl.acm.org/doi/10.1145/...
as long as CA is banning loud commercials on streaming services, can we also ban loud logo soundmarks too (e.g., Netflix)? i've taught myself to mute them, but they still occasionally slip through TA-DUMMMMM
October 9, 2025 at 8:28 AM
as long as CA is banning loud commercials on streaming services, can we also ban loud logo soundmarks too (e.g., Netflix)? i've taught myself to mute them, but they still occasionally slip through TA-DUMMMMM
When a chatbot’s words *are* its interface, how do we see the aesthetics of engagement they activate—and the politics they conceal? New article w/ @spold.bsky.social
dl.acm.org/doi/10.1145/...
dl.acm.org/doi/10.1145/...
September 15, 2025 at 6:09 PM
When a chatbot’s words *are* its interface, how do we see the aesthetics of engagement they activate—and the politics they conceal? New article w/ @spold.bsky.social
dl.acm.org/doi/10.1145/...
dl.acm.org/doi/10.1145/...
New paper w/ @spold.bsky.social. We introduce new methods to critique linguistic interfaces of AI chatbots, theorize one of ChatGPT's central engagement mechanics as the "praise/prompt envelope," and preview artworks that attune users to see its aesthetics of engagement dl.acm.org/doi/10.1145/...
August 25, 2025 at 5:46 PM
New paper w/ @spold.bsky.social. We introduce new methods to critique linguistic interfaces of AI chatbots, theorize one of ChatGPT's central engagement mechanics as the "praise/prompt envelope," and preview artworks that attune users to see its aesthetics of engagement dl.acm.org/doi/10.1145/...
we need public social media
June 6, 2025 at 1:04 AM
we need public social media
two billionaires, each with their own personal tech platforms—online forums where agreeing with the owner is the baseline condition for visibility—fight in public while the world watches. welcome to social media in 2025.
June 6, 2025 at 12:31 AM
two billionaires, each with their own personal tech platforms—online forums where agreeing with the owner is the baseline condition for visibility—fight in public while the world watches. welcome to social media in 2025.
ChatGPT's palpably excessive praise over the last week isn't an error or misconfig—it's a deliberate strategy by OpenAI to fine tune praise as an engagement tactic. They've been slowly turning that dial for 2 yrs and only now did users start to notice (I've been recording its praise the whole time)
April 29, 2025 at 6:49 AM
ChatGPT's palpably excessive praise over the last week isn't an error or misconfig—it's a deliberate strategy by OpenAI to fine tune praise as an engagement tactic. They've been slowly turning that dial for 2 yrs and only now did users start to notice (I've been recording its praise the whole time)
if Facebook really cared about giving its users a no algo feed, it would make their new friends feed the *default*
also, it’s a sign of how bad Facebook’s become that they had to build a new friends feed at all
also, it’s a sign of how bad Facebook’s become that they had to build a new friends feed at all
March 28, 2025 at 2:28 AM
if Facebook really cared about giving its users a no algo feed, it would make their new friends feed the *default*
also, it’s a sign of how bad Facebook’s become that they had to build a new friends feed at all
also, it’s a sign of how bad Facebook’s become that they had to build a new friends feed at all
TikTok has increased frequency of ads on my For You to 1 ad for every 2-3 videos.
February 18, 2025 at 11:18 PM
TikTok has increased frequency of ads on my For You to 1 ad for every 2-3 videos.
there's no way to chat with deepseek's web interface without giving up your data for model training or any other purpose. while i don't presume openAI/anthropic/etc won't violate the opt-out agreement we have (as a paying user), at least there *is* an agreement
January 24, 2025 at 5:24 AM
there's no way to chat with deepseek's web interface without giving up your data for model training or any other purpose. while i don't presume openAI/anthropic/etc won't violate the opt-out agreement we have (as a paying user), at least there *is* an agreement
with tiktok back up, i’m still stuckinthescroll.com
Stuck in the Scroll
A compulsive scrolling monitor for TikTok.
stuckinthescroll.com
January 19, 2025 at 11:13 PM
with tiktok back up, i’m still stuckinthescroll.com
My social platform Minus will never be sold, never monetize your data, and never shutdown for anyone anywhere (unless it breaks and I'm unable to fix it). It also won't manipulate your psyche to extract more data or keep you endlessly scrolling or clicking. minus.social
Minus
A finite social network where you get 100 posts—for life
minus.social
January 19, 2025 at 8:25 AM
My social platform Minus will never be sold, never monetize your data, and never shutdown for anyone anywhere (unless it breaks and I'm unable to fix it). It also won't manipulate your psyche to extract more data or keep you endlessly scrolling or clicking. minus.social
if tiktok goes dark in the US, i expect there will be a non-zero number of humans who look up from their phones for the first time in 5 years
January 19, 2025 at 1:30 AM
if tiktok goes dark in the US, i expect there will be a non-zero number of humans who look up from their phones for the first time in 5 years
Whether or not TikTok goes dark in the US this Sunday, the looming shutdown exposes just how fragile and controlled our access to software has become.
January 17, 2025 at 6:53 PM
Whether or not TikTok goes dark in the US this Sunday, the looming shutdown exposes just how fragile and controlled our access to software has become.
Whether or not TikTok gets a reprieve, what the looming ban makes most palpable is how important it is to get (what feels to many like) public culture out of private big tech platforms.
January 16, 2025 at 10:36 PM
Whether or not TikTok gets a reprieve, what the looming ban makes most palpable is how important it is to get (what feels to many like) public culture out of private big tech platforms.
Zuckerberg says many laughable things, but saying he will “intensely” focus on “the future of social media” is certainly high on the list. His approach—for the last 15 years at least—has been to hold social media in stasis, to freeze the idea for almost everyone. So little change, so much copying.
January 15, 2025 at 3:44 AM
Zuckerberg says many laughable things, but saying he will “intensely” focus on “the future of social media” is certainly high on the list. His approach—for the last 15 years at least—has been to hold social media in stasis, to freeze the idea for almost everyone. So little change, so much copying.
Zuck only cares about one thing. It's been consistent since day 1. Everything else is secondary vimeo.com/333795857
ORDER OF MAGNITUDE
As the founder and CEO of the world’s largest social media corporation, what does Mark Zuckerberg think about? While we get clues from his posts on Facebook…
vimeo.com
January 10, 2025 at 3:09 AM
Zuck only cares about one thing. It's been consistent since day 1. Everything else is secondary vimeo.com/333795857
Zuck must believe cranking up mis/disinfo will benefit engagement. I expect he’s right—lies and obfuscations will induce impassioned reaction by those who believe and those who don’t. In fact, I expect the engagement boost from his announcement is already paying off.
January 9, 2025 at 10:10 AM
Zuck must believe cranking up mis/disinfo will benefit engagement. I expect he’s right—lies and obfuscations will induce impassioned reaction by those who believe and those who don’t. In fact, I expect the engagement boost from his announcement is already paying off.
Zuckerberg has spent the last 20 years conditioning users to become machinic in their online interactions. So, flooding his platforms with AI machine “friends” must feel like a next logical step towards further automations of user behavior.
www.rollingstone.com/culture/cult...
www.rollingstone.com/culture/cult...
Facebook and Instagram to Unleash AI-Generated 'Users' No One Asked For
Meta says it will integrate AI-generated characters into Facebook and Instagram with their own user accounts.
www.rollingstone.com
January 3, 2025 at 1:09 AM
Zuckerberg has spent the last 20 years conditioning users to become machinic in their online interactions. So, flooding his platforms with AI machine “friends” must feel like a next logical step towards further automations of user behavior.
www.rollingstone.com/culture/cult...
www.rollingstone.com/culture/cult...
I want a t-shirt with one of Claude's favorite sentences-->
Tip: Long chats cause you to reach your usage limits faster.
Tip: Long chats cause you to reach your usage limits faster.
December 25, 2024 at 8:46 AM
I want a t-shirt with one of Claude's favorite sentences-->
Tip: Long chats cause you to reach your usage limits faster.
Tip: Long chats cause you to reach your usage limits faster.
OpenAI's Sora ignores relevant data preferences you've already set with ChatGPT and instead defaults to sharing your data with the model and publishing your videos for everyone. It should import those preferences, not reset them.
December 17, 2024 at 9:43 PM
OpenAI's Sora ignores relevant data preferences you've already set with ChatGPT and instead defaults to sharing your data with the model and publishing your videos for everyone. It should import those preferences, not reset them.
automatic hi beams (brights) on cars are designed for a world in which there are no pedestrians. they presume that the absence of opposing headlights means it’s all clear to blind any living thing in their path. sadly they also fail regularly for cars too, so everyone in/out gets blinded.
December 17, 2024 at 9:01 PM
automatic hi beams (brights) on cars are designed for a world in which there are no pedestrians. they presume that the absence of opposing headlights means it’s all clear to blind any living thing in their path. sadly they also fail regularly for cars too, so everyone in/out gets blinded.
chatgpt sure is slow tonight. i like it
December 16, 2024 at 10:30 AM
chatgpt sure is slow tonight. i like it
chatgpt’s default voice performance style has this enthusiastic advertising/marketing aesthetic to it. at times i almost feel like the answers are performed in the form of a radio commercial. then i notice i feel bad when I ask it to moderate that feel—and it (kind of) does.
December 15, 2024 at 11:37 PM
chatgpt’s default voice performance style has this enthusiastic advertising/marketing aesthetic to it. at times i almost feel like the answers are performed in the form of a radio commercial. then i notice i feel bad when I ask it to moderate that feel—and it (kind of) does.