Opens profile photo
Follow
Click to Follow pranesh
Pranesh Prakash
@pranesh
prāṇēś prakāś Law+policy consultant: #a2k #privacy #FoE #AI #infosec #FOSS #a11y #NetNeutrality zhōngyōng & anēkāntavāda. Past: / Hire me!
Madras, Tamil Nadu, Indiaorcid.org/0000-0002-5368…Joined July 2007

Pranesh Prakash’s Tweets

This is big. No #embargoes. No #APCs. "The #EU is ready to agree that immediate #OpenAccess to…publicly funded research should become the norm, w/o authors having to pay fees & that the bloc should support #nonprofit scholarly publishing models."
3
128
Lecture by Prof. Samuelson on #AI & #copyright (looking at both output & input). This is in the US context, but is useful globally since it's not abt 'fair use'. She notes that © law could be an "existential threat" to generative AI. (via )
3
"Are these the last convulsions of an incoherent order which was hollowed out long ago by the decline of Christianity and the end of Britain as a world power?" Good article by @.
Screenshot from https://unherd.com/thepost/todays-coronation-is-a-wonderful-contradiction-in-terms/
Quote Tweet
It's hard to describe how bizarre today's coronation will be, even by the standards of monarchies. But you will remember where you were when you saw it. Me for @unherd. unherd.com/thepost/todays
2
If *copying* copyrighted works for search engines isn't seen as copyright violation, then *training* on copyrighted works ought not be seen as a copyright violation either. It's clearly "transformative use"/fair use, and should be seen as a fair dealing. #copyright #AI
Quote Tweet
Hey @MosaicML. If you fine-tune on books3 full of 197,000 Copyrighted works, are you sure you have legal rights to relicense the model as commercial? Meta has been unable to license LLaMa because they know © law, not because they want to keep it private... FYI @AuthorsGuild twitter.com/vitaliychiley/…
Show this thread
Image
2
7
This interview, while purportedly about important issues around AI ethics and safety, seems to boil down to Google office politics and tech policy personality cults.
Quote Tweet
My full interview with @mer__edith, about what's missing from Geoffrey Hinton's media tour, how AI doomers put us into a "trance," and why labor organizing is still the most powerful check against Big Tech: fastcompany.com/90892235/resea
Show this thread
1
"Assigned male/female at birth" (in any context other than DSDs) is among the dumbest reality-distorting phrases that I've ever seen. Journalists and academics using such verbiage reveal themselves to be thoroughly unserious and disconnected from reality.
Quote Tweet
Replying to @JamesSurowiecki @wesyang and @SwipeWright
"It's a girl, since it has a vagina," is no more an assignment than, "it's a zebra, since it looks like a donkey with stripes." Both are inferences abt the whole based on parts. Neither, contra Butler & Fausto-Sterling, is a normative act (an "assignment"), neither is illocution.
5
The combo of AI-related concerns about disinformation + online safety + copyright all converge on common set of regulations: Identity mandates / watermarking / expanded platform liability / banning most kids online. It’ll be a remarkably sweeping govt information control regime.
2
19
Show this thread
Michio Kaku has been going batty for a while now. This thread is both fascinating and sad.
Quote Tweet
@MichioKaku: "Quantum computers can act as a fact checker on ChatGPT." @JoeRogan: "It's going to be able to discern what's real and what's not real?" @MichioKaku: "Yes, and even gradations of what's true. It can give you detailed understanding of what is partially correct."
Show this thread
Embedded video
1:22
144.9K views
2
4
I offer a brief comment as a "cyber expert" (??) on the current rhetoric on AI risks and US White House's actions to promote responsible AI innovation. whitehouse.gov/briefing-room/
Quote Tweet
WHITE HOUSE SUMMONS BIG TECH BOSSES Alarm Bells Over Artificial Intelligence "People are over estimating the #ArtificialIntelligence risk and under estimating the benefit:" @pranesh, Cyber Expert to @radhika1705
Embedded video
2:22
189 views
1
The soft bigotry of low expectations: "underrepresented minority students" are "systematically shortchange[d]" by teaching through lectures; "multiple-choice test[s]" "fail to promote social justice principles". This is harmful for the very people it is meant to aid.
Quote Tweet
Black students don't learn from lectures. So says UNC's Dept. of Allied Health Sciences (@UNC_HealthSci). UNC's new DEI plan will require faculty to "integrate social justice content into their teaching," apparently without lecturing. @JohnDSailer: nas.org/blogs/article/
Image
4
As per Mozilla's content policy, "posts … that contain derogatory and/or hateful language" about age, sex, national origin, etc., won't be allowed. So would "OK, Boomer" or "All men are rapists, and that's all they are" be considered "derogatory" on the basis of age, sex? 2/2
2
7
Show this thread
Mozilla is quite clear that liberalism — the idea that we should respect and accept behaviour or opinions different from our own preferences — is harmful and antithetical to "human dignity, inclusion, security, individual expression (??) and collaboration". 1/2

You’ll notice a big difference in our content moderation approach compared to other major social media platforms. We’re not building another self-declared “neutral” platform. We believe that far too often, “neutrality” is used as an excuse to allow behaviors and content that’s designed to harass and harm those from communities that have always faced harassment and violence. Our content moderation plan is rooted in the goals and values expressed in our Mozilla Manifesto — human dignity, inclusion, security, individual expression and collaboration. We understand that individual expression is often seen, particularly in the US, as an absolute right to free speech at any cost. Even if that cost is harm to others. We do not subscribe to this view. We want to be clear about this. We’re building an awesome sandbox for us all to play in, but it comes with rules governing how we engage with one another. You’re completely free to go elsewhere if you don’t like them.
3
8
Show this thread
Not an easy choice. I think there's a wider variety of tastes in American Chinese than in Indian Chinese. But overall, I think Indian Chinese wins out.
Quote Tweet
Indianized Chinese cuisine vs. Americanized Chinese cuisine, who do you prefer?
Image
Image
1
1