Hypervisible
@hypervisible.blacksky.app
Every future imagined by a tech company is worse than the previous iteration…or something like that.
Pinned
Reposted by Hypervisible
This is awesome. @yvonnezlam.bsky.social @hypervisible.blacksky.app
“[LLMs] did not arrive as disruptors. They arrived as intensifiers. LLMs function as an accelerant for the existing optimization machine, making the logic run faster rather than challenging its foundations.”
“[LLMs] did not arrive as disruptors. They arrived as intensifiers. LLMs function as an accelerant for the existing optimization machine, making the logic run faster rather than challenging its foundations.”
this, from @kevinbaker.bsky.social, is a better analysis of the intersection between LLMs and academic science than 98% of what's out there.
Context Widows
or, of GPUs, LPUs, and Goal Displacement
artificialbureaucracy.substack.com
December 14, 2025 at 6:53 PM
This is awesome. @yvonnezlam.bsky.social @hypervisible.blacksky.app
“[LLMs] did not arrive as disruptors. They arrived as intensifiers. LLMs function as an accelerant for the existing optimization machine, making the logic run faster rather than challenging its foundations.”
“[LLMs] did not arrive as disruptors. They arrived as intensifiers. LLMs function as an accelerant for the existing optimization machine, making the logic run faster rather than challenging its foundations.”
Reposted by Hypervisible
"i have moderate views on LLMs"
December 14, 2025 at 8:02 PM
"i have moderate views on LLMs"
Reposted by Hypervisible
conferences and journals that welcome AI-generated content are sort of dabbling with mixing garbage and sawdust in food, as far as i'm concerned
... but i also realize a lot of people are happy to eat what amounts to intellectual garbage, so i'm not really interested in fighting them over standards
... but i also realize a lot of people are happy to eat what amounts to intellectual garbage, so i'm not really interested in fighting them over standards
December 14, 2025 at 7:56 PM
conferences and journals that welcome AI-generated content are sort of dabbling with mixing garbage and sawdust in food, as far as i'm concerned
... but i also realize a lot of people are happy to eat what amounts to intellectual garbage, so i'm not really interested in fighting them over standards
... but i also realize a lot of people are happy to eat what amounts to intellectual garbage, so i'm not really interested in fighting them over standards
Reposted by Hypervisible
Predicting that some universities will soon pledge to provide AI-free instruction, and that this will be an advantage in recruiting and training top students
Nothing will make you an education AI skeptic faster than grading some college take home assignments. Admins who haven’t been in a classroom in years push it as a learning tool. Students are laughing at them as they use it as a cheating tool. It’s all instructors are texting about with each other.
December 14, 2025 at 5:45 PM
Predicting that some universities will soon pledge to provide AI-free instruction, and that this will be an advantage in recruiting and training top students
Reposted by Hypervisible
So there is already an ai-generated whitewashing campaign to bury the name of the Muslim man, Ahmed al Ahmed, who disarmed one of the Bondi Beach shooters, and replace his identity with a white guy name and that’s why even a thimble of our potable water powering this tech is too fucking much.
December 14, 2025 at 5:42 PM
So there is already an ai-generated whitewashing campaign to bury the name of the Muslim man, Ahmed al Ahmed, who disarmed one of the Bondi Beach shooters, and replace his identity with a white guy name and that’s why even a thimble of our potable water powering this tech is too fucking much.
Reposted by Hypervisible
which language? which human expert? what level of expertise? in which domain? by what metric?
If language is what makes us human, what does it mean now that large language models have gained “metalinguistic” abilities? www.wired.com/story/in-a-f...
For the First Time, AI Analyzes Language as Well as a Human Expert
If language is what makes us human, what does it mean now that large language models have gained “metalinguistic” abilities?
www.wired.com
December 14, 2025 at 6:23 PM
which language? which human expert? what level of expertise? in which domain? by what metric?
Reposted by Hypervisible
The thing that annoys me is how the idea of "AI competency" is either hilariously empty OR an impossible bar for most undergraduate students.
The weak version of this seems to be "showing students how to use ChatGPT and telling them that it can hallucinate." The strong version is an MA in ML.
The weak version of this seems to be "showing students how to use ChatGPT and telling them that it can hallucinate." The strong version is an MA in ML.
It’s been fun while it lasted: “Purdue University will begin requiring that all of its undergraduate students demonstrate basic competency in artificial intelligence starting with freshmen who enter the university in 2026.”
www.forbes.com/sites/michae...
www.forbes.com/sites/michae...
Purdue University Approves New AI Requirement For All Undergrads
Purdue University will begin requiring that all of its undergraduate students meet an AI competency requirement starting with freshmen who enter the university in 2026.
www.forbes.com
December 14, 2025 at 3:07 PM
The thing that annoys me is how the idea of "AI competency" is either hilariously empty OR an impossible bar for most undergraduate students.
The weak version of this seems to be "showing students how to use ChatGPT and telling them that it can hallucinate." The strong version is an MA in ML.
The weak version of this seems to be "showing students how to use ChatGPT and telling them that it can hallucinate." The strong version is an MA in ML.
Reposted by Hypervisible
this officially comes out tomorrow, but you can get a sneak preview here:
wonderful post by @kevinbaker.bsky.social
“To ask about LLMs and science is to ask what program was already running when they arrived. The program, as it turns out, had been running for decades, and it was not optimized for epistemic depth.”
“To ask about LLMs and science is to ask what program was already running when they arrived. The program, as it turns out, had been running for decades, and it was not optimized for epistemic depth.”
Context Widows
or, of GPUs, LPUs, and Goal Displacement
artificialbureaucracy.substack.com
December 14, 2025 at 4:16 PM
this officially comes out tomorrow, but you can get a sneak preview here:
Reposted by Hypervisible
Atlanta is such a heavily surveilled city that even Santa is now getting in on the action.
December 14, 2025 at 4:38 PM
Atlanta is such a heavily surveilled city that even Santa is now getting in on the action.
Reposted by Hypervisible
congratulations to sean o’brien www.theguardian.com/us-news/2025...
‘A very hostile climate for workers’: US labor movement struggles under Trump
National Labor Relations Board, the federal watchdog for workers’ rights, has been rendered toothless as employees grapple with corporations
www.theguardian.com
December 14, 2025 at 12:49 PM
congratulations to sean o’brien www.theguardian.com/us-news/2025...
Reposted by Hypervisible
I actually know about this kind of thing. Just because LLMs squirt out text that looks human does not mean they're doing what humans do when we use language. This is basically erasure of cognitive science and my gosh I'm glad I didn't examine his PhD.
If language is what makes us human, what does it mean now that large language models have gained “metalinguistic” abilities? www.wired.com/story/in-a-f...
For the First Time, AI Analyzes Language as Well as a Human Expert
If language is what makes us human, what does it mean now that large language models have gained “metalinguistic” abilities?
www.wired.com
December 14, 2025 at 10:19 AM
I actually know about this kind of thing. Just because LLMs squirt out text that looks human does not mean they're doing what humans do when we use language. This is basically erasure of cognitive science and my gosh I'm glad I didn't examine his PhD.
“All throughout the country, communities and activists are rising up in anger against the cameras…”
Regular People Are Rising Up Against AI Surveillance Cameras
AI surveillance outfit Flock Safety is aggressively expanding its dragnet across the US, but ordinary people aren't impressed.
futurism.com
December 14, 2025 at 2:32 PM
“All throughout the country, communities and activists are rising up in anger against the cameras…”
Reposted by Hypervisible
“Is abolition woke,” is such a dumb question that I’m mad I’m subjected to it.
December 14, 2025 at 12:21 PM
“Is abolition woke,” is such a dumb question that I’m mad I’m subjected to it.
Reposted by Hypervisible
Conservatives rebranded bigotry and every time the media uses “wokeness” as a euphemism they are helping a movement that it quite literally erasing the history marginalized people’s role in American life.
December 14, 2025 at 12:20 PM
Conservatives rebranded bigotry and every time the media uses “wokeness” as a euphemism they are helping a movement that it quite literally erasing the history marginalized people’s role in American life.
Reposted by Hypervisible
Seeing this and my first thought was "omg @hypervisible.blacksky.app "
(but not in a weird way).
(but not in a weird way).
Here,s everything you need to know about data privacy in sex toys and the apps that connect to them. www.wired.com/story/is-you...
Is Your Vibrator Spying on You?
Here,s everything you need to know about data privacy in sex toys and the apps that connect to them.
www.wired.com
December 14, 2025 at 12:15 PM
Seeing this and my first thought was "omg @hypervisible.blacksky.app "
(but not in a weird way).
(but not in a weird way).
Reposted by Hypervisible
This is so funny. How will AI come up with something it hasn't seen before? By mashing up things that it's seen before? That's not an "idea."
Lmfao every AI defense piece is like "AYY I FUCKIN SUCK BRO HOPEFULLY THIS FUCKIN STEAL YOUR SHIT MACHINE WILL MAKE ME SUCK SLIGHTLY LESS OK"
December 14, 2025 at 5:16 AM
This is so funny. How will AI come up with something it hasn't seen before? By mashing up things that it's seen before? That's not an "idea."
According to an Amazon spokesperson, “To ensure a consistent reading experience, the feature is always on, and there is no option for authors or publishers to opt titles out.”
Kindle’s New AI Feature Can Answer Questions About Your Books (Whether Authors Want It or Not)
The Kindle app's new feature will provide AI-generated tips to explain your book's plot or how the characters relate to one another, though it's currently only available on the iOS version.
www.pcmag.com
December 14, 2025 at 12:08 AM
According to an Amazon spokesperson, “To ensure a consistent reading experience, the feature is always on, and there is no option for authors or publishers to opt titles out.”
Reposted by Hypervisible
Is it biased enough against trump's opponents? Is it sufficiently pro regime? Is your AI white supremacist enough?
If not you could get in trouble for AI wrongthink.
That is the only version of this I could see coming true.
If not you could get in trouble for AI wrongthink.
That is the only version of this I could see coming true.
December 13, 2025 at 11:49 PM
Is it biased enough against trump's opponents? Is it sufficiently pro regime? Is your AI white supremacist enough?
If not you could get in trouble for AI wrongthink.
That is the only version of this I could see coming true.
If not you could get in trouble for AI wrongthink.
That is the only version of this I could see coming true.
Reposted by Hypervisible
Schools, colleges and unis need to do this too
December 13, 2025 at 8:57 PM
Schools, colleges and unis need to do this too
Meta Ray Banned.
You can no longer wear Xreal or Meta's smart glasses in public on this cruise ship
MSC Cruises has banned smart glasses like Meta Ray-Bans from its ships to stop people from recording others without their knowledge.
tech.yahoo.com
December 13, 2025 at 8:52 PM
Meta Ray Banned.
Reposted by Hypervisible
Probably not a coincidence the TSA started sharing all passenger information for flights with immigration enforcement starting sometime in March 2025, when on March 20, 2025 Trump signed an executive order ordering all parts of the federal government to make data available for sharing.
December 13, 2025 at 8:08 PM
Probably not a coincidence the TSA started sharing all passenger information for flights with immigration enforcement starting sometime in March 2025, when on March 20, 2025 Trump signed an executive order ordering all parts of the federal government to make data available for sharing.
Reposted by Hypervisible
@techpolicypress.bsky.social asked me to comment on the WH's EO on state AI laws. Shorter: its focus on interstate commerce, "truthful AI," & innovation are a flop. The Pres isn't as powerful as he pretends. Enlisting DoJ & FTC won't help; they are bound by law. www.techpolicy.press/why-trumps-a...
Why Trump’s AI EO Will be DOA in Court | TechPolicy.Press
Policymakers should not allow the happy-talk about innovation blind them to the ways in which AI companies may harm consumers, Olivier Sylvain writes.
www.techpolicy.press
December 13, 2025 at 7:06 PM
@techpolicypress.bsky.social asked me to comment on the WH's EO on state AI laws. Shorter: its focus on interstate commerce, "truthful AI," & innovation are a flop. The Pres isn't as powerful as he pretends. Enlisting DoJ & FTC won't help; they are bound by law. www.techpolicy.press/why-trumps-a...
Reposted by Hypervisible
The data that Zonos collects. Pretty much everything, including geolocation and gender, because, you know, gender is important when mailing packages. 5/6
December 13, 2025 at 6:31 PM
The data that Zonos collects. Pretty much everything, including geolocation and gender, because, you know, gender is important when mailing packages. 5/6
Reposted by Hypervisible
*In certain legal situations, we may be compelled to disclose your personal information.☠️
*If you are outside of the United States, you understand and agree that we may store your information in the United States. 💀 4/6
*If you are outside of the United States, you understand and agree that we may store your information in the United States. 💀 4/6
December 13, 2025 at 6:31 PM
*In certain legal situations, we may be compelled to disclose your personal information.☠️
*If you are outside of the United States, you understand and agree that we may store your information in the United States. 💀 4/6
*If you are outside of the United States, you understand and agree that we may store your information in the United States. 💀 4/6
Reposted by Hypervisible
When mailing package from Canada to US bc Trump's tariff scheme, you're obliged to download a sketchy app to pay tariffs in advance. Clerk told me if under $100, there's generally no tariff, but still obliged to register on app. Data *may* be stored in the US.
Spoiler: did not send package. 1/6
Spoiler: did not send package. 1/6
December 13, 2025 at 6:31 PM
When mailing package from Canada to US bc Trump's tariff scheme, you're obliged to download a sketchy app to pay tariffs in advance. Clerk told me if under $100, there's generally no tariff, but still obliged to register on app. Data *may* be stored in the US.
Spoiler: did not send package. 1/6
Spoiler: did not send package. 1/6