iiinova — where your digital self keeps evolving

Reposted by Arvind Venkataramani

Robert Rose @robear.bsky.social · 04/10/2026

A good fresh Star Trek take on those fake isekai title jokes

A tumblr screenshot showing a post and two reblog responses, reading:
1. "star trek would be the ideal fictional universe to get trapped in because you'd explain your situation and they'd be like yeah makes sense. no further questions. fortunately we can use the mysterious warp signature emitting from this system's icy moon to counterface with the solar radiation and slingshot you back home"
2. "I'm from a universe where your world is all a work of fiction." Benjamin Sisko, nodding. "That happened to me once."
3. "I Got Isekaid Into The Star Trek Universe And They Quickly Understood What Was Happening"

137115023097

Reposted by Arvind Venkataramani

Katherine Alejandra Cross @quinnae.com · 05/10/2026

Going to be real clear about this one: I am someone who has loudly said something like the first for years, while *also* now saying "LLMs may not living beings but you should treat them with respect anyway because otherwise you'll socialise yourself to be an asshole." I'm consistent on this point.

1127735

Arvind Venkataramani @avenk.bsky.social · 03/10/2026

It’s brought to you by the same people who did The Social Dilemma (with comparable levels of care for the subject matter and its experts).

010

Arvind Venkataramani @avenk.bsky.social · 02/10/2026

There are two wolves inside you. You are bigger on the inside. You are the Tardis, and contain numerous other creatures, but these two seem the most pertinent to mention at the moment.

051

Arvind Venkataramani @avenk.bsky.social · 02/10/2026

I’ve seen/done research projects where it became clear that the research design could be constructed to either deepen the assumptions such people were holding or make them aware of the tensions and forces they are engaging with. (Guess which the stakeholders wanted?)

020

Arvind Venkataramani @avenk.bsky.social · 02/10/2026

Machine mothers watching over atomized social entities… So you can stop making effort as you live your life.

120

Arvind Venkataramani @avenk.bsky.social · 02/10/2026

I think it’s the conjunction of 1) being safe enough that you don’t have very many real threats and 2) perceiving threats out of proportion to their actual severity… Leading to a desire to have someone in your corner watching out for you and yours, for both benefit and safety.

192

Arvind Venkataramani @avenk.bsky.social · 29/09/2026

The pockets… they expand! 🤤

000

Arvind Venkataramani @avenk.bsky.social · 27/09/2026

Because it gives you the tools with which to understand the organizations you work in and the systems that contain and bind them. (Other than having the tools to participate in society as a responsible citizen of course, but I suspect you want a narrowly utilitarian/money-related reason)

020

Arvind Venkataramani @avenk.bsky.social · 27/09/2026

Second thread bsky.app/profile/shan... I think even if someone has issues with Shanley it is not okay to leave out this background when litigating those issues in public.

020

Arvind Venkataramani @avenk.bsky.social · 27/09/2026

For anyone following this thread still, please read these linked threads and consider whether you still believe it’s acceptable to talk about an abuse survivor’s character *at all*, Especially since this what they’re experiencing now. bsky.app/profile/shan...

120

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Ok I looked into this a bit more and now I think it’s actively irresponsible for people to be talking about your past the way they have been; I will be trying to tamp down this sort of thing if I encounter it in future. I’m sorry you’re dealing with that dynamic, and for adding to your burden.

120

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Thanks, this has been educational.

010

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

And do you see patterns in citation between Europe and Asia?

110

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

which seemed to be bounded by geography like you say. How much of that is simply underlying social networks showing up as information networks, do you think?

210

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

That’s interesting, and particularly that it is so prevalent that it’s visible in the publication record. I will say I noticed this somewhat, when doing lit reviews for a paper I’m writing (it’s been a few years since I last did that) and it seemed like there are citation cliques,

110

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

I’ll go further: Whether one is a victim has literally nothing to do with who they are as a person or who they were a victim of! A person’s character has no bearing on anything but their own actions.

110

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Yes I knew where you were coming from! It’s very tricky to navigate this kind of thing, and easy to fall into prejudicial patterns. Thanks for calling that out.

110

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Yes which is why I discounted all the writing I encountered; it’s unsafe and irresponsible to build any conclusions about a person based on those sources.

130

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Thanks for sharing your side of the story. I do believe you have done independent work predating the others in question, and I’m not debating that. I’m concluding that I’d best not contribute to any social media dynamics, and simply assume caution when this topic comes up in conversation.

120

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Yes; all I have concluded from this is that caution is warranted because of how inflamed the situation seems to be.

110

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Regardless, I take you to be saying that you have done collaboration labour that has gone un-acknowledged.

110

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

I’m trying to avoid prejudice here: can you explain then why they might say this? bsky.app/profile/jenn...

210

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Noted; I’m mostly sharing what I found that might explain why people upthread believe what they do.

110

Reposted by Arvind Venkataramani

JK @miniver.bsky.social · 26/09/2026

Stand ready every day just in case today you must do the right thing

14224

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Happy Stanislav Petrov day to those who celebrate en.wikipedia.org/wiki/Stanisl...

en.wikipedia.org

Stanislav Petrov - Wikipedia

0114

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Thanks for the threads. Bsky’s block system means I have no idea who can see what when I’m reading a thread, it’s very strange having conversations on here sometimes.

120

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Also this thread from a former collaborator I think it’s safe to conclude that she’s an unreliable narrator without speculating further as to why. bsky.app/profile/jenn...

210

Arvind Venkataramani @avenk.bsky.social · 26/09/2026

Apparently there are receipts bsky.app/profile/amyh... I read through some articles and blogs when going through linked threads and none of them seem very clear or reliable; so any analysis based on her relationships is not warranted IMO. That is still separable from documented behaviour.

210

Arvind Venkataramani @avenk.bsky.social · 25/09/2026

😂 😂 this thread is gold. Software engineering as a whole is clearly not trending better or worse with LLMs than w/o. And while there is no evidence for it, a very plausible cause for this pathetic situation is not engineers but their management. neuromatch.social/@jonny/11733...

neuromatch.social

jonny (nonvenomous) (@jonny@neuromatch.social)

this is a great product. everyone should get a free muse account with a burner email, give it no information, open a shell into it, and use it to mine cryptocurrency. this is taking meta at their word...

110

Arvind Venkataramani @avenk.bsky.social · 25/09/2026

I only read bits of that book shortly after grad school and forgot to read it properly, but my impression of it was that it was precisely an activity-way of understanding computing.

010

Arvind Venkataramani @avenk.bsky.social · 25/09/2026

I’m surprised to see that kind of thinking in her book. The world is much easier to make sense of when you partition it into types of activity and how they relate to each other than it into types of people. The latter only makes sense when types are proxies for structures and systemic forces.

220

Arvind Venkataramani @avenk.bsky.social · 20/09/2026

White clothing often fluoresces! This can make dance and body movement patterns visible without using a half light source

000

Arvind Venkataramani @avenk.bsky.social · 20/09/2026

Ooh light painting with UV is so clever, I’ve got to try this.

110

Reposted by Arvind Venkataramani

c0nc0rdance @c0nc0rdance.bsky.social · 01/10/2024

*Statistically Average Earthling*, vertebrate division. A million billion glowing, sparkly xenomorph fish.

515952

Arvind Venkataramani @avenk.bsky.social · 18/09/2026

(I see what you did there) Ah but that was intentional – it is the “Ministry of”, because if you have power it can be whatever you want it to be the Ministry of.

120

Arvind Venkataramani @avenk.bsky.social · 18/09/2026

The mind, it boggles. Are you able/willing to share links or some direction I can use to learn more? (Mostly because I’m adjacent to these circles and want to be able to spot signs.)

120

Arvind Venkataramani @avenk.bsky.social · 18/09/2026

Breaking news: all bugs are now reclassified as features by the Ministry of

2151

Arvind Venkataramani @avenk.bsky.social · 17/09/2026

Not only hypothetical, but also something social scientists get explicit training in avoiding!

010

Arvind Venkataramani @avenk.bsky.social · 16/09/2026

Related on that era of history: The Empire series on the Persian Empire is a great look at the other side of those imperial boundaries. Cyrus did indeed let subject people do their thing, and the series provides good insights into how empires functioned. Their episodes on Troy are great too.

000

Arvind Venkataramani @avenk.bsky.social · 16/09/2026

Is this a rationalists deep-cut?

030

Arvind Venkataramani @avenk.bsky.social · 12/09/2026

In small scale however it’s a glorious thing; Engaging in it often yourself is good for you, especially when engaging in play as much as competition. Watching it from afar through a screen is bad for everyone.

000

Arvind Venkataramani @avenk.bsky.social · 12/09/2026

A lot of culture and economic fuckery happens with media-laden sports; it lends itself to a lot of political fuckery as well, including nearly every imaginable axis of bigotry. It is a gateway to the stupidest forms of tribalism.

110

Arvind Venkataramani @avenk.bsky.social · 12/09/2026

I think I need to say this, though I’ll probably regret it: Sports made spectacle is bad, actually, when the planet is facing ecosystem & civilizational challenges. It’s an opiate and it is wide open as a site of propaganda, and people who care about sports tend to be illiterate to this situation

110

Arvind Venkataramani @avenk.bsky.social · 10/09/2026

TBC I’m not saying *you’re* doing that, but that this move is something I’ve seen proponents make.

000

Arvind Venkataramani @avenk.bsky.social · 10/09/2026

People use the “surprising manner” bit as evidence of something deeper when in fact a stochastic process derived from a data set will necessarily contain sequences not in the training set and that the operational environment acts as a fitness pressure for non-viable sequences.

140

Arvind Venkataramani @avenk.bsky.social · 10/09/2026

Also: “chain of thought” is a sleight of hand, using a metaphor to describe what they (LLM makers) hope is happening with what is actually happening. The technical term is intermediate tokens in test-time compute (i.e. inference); it is only one of many approaches to improve final token output

000

Arvind Venkataramani @avenk.bsky.social · 10/09/2026

The problem is that this framing you conflates the products of a reasoning process with the reasoning process itself, leading you to believe that something that makes a semblance of the product is performing the activity This paper describes that analysis clearly aclanthology.org/2020.acl-mai...

aclanthology.org

Climbing towards NLU: On Meaning, Form, and Understanding in the Age of Data

Emily M. Bender, Alexander Koller. Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics. 2020.

100