Chris Kubick

so called sounds from the life of chris kubick, mastodon: @[email protected]

[Yi Ma @YiMaTweets]

Interestingly, the last chapter of my An Invitation to 3D Vision book is all about why symmetry makes 3D perception easy. Seeing abstract patterns in data seem to be human's unique ability. Does this ability to abstract patterns come from certain extreme form of compression? Maybe, or maybe a little more than that.

[Chris Kubick @SoCalledSound]

@suchenzang @Sentdex can we get an option for lol please

[Chris Kubick @SoCalledSound]

@matdryhurst @neilturkewitz @benjaminjriley I honestly just wish people would stop using the term AI. Playing with the probabilities of language or model weightings or machine vision is just not the same thing as asking questions to a corporate chatbot. Unfortunately Rubin just comes off as another oneshotted boomer

[Max Carpenter @solexyoghurt]

here are nine killer IDM records that I'm sure most laylisteners don't know (if you do you get a gold star ⭐️). no Warp stuff, and, for the most part, no big names:

[Chris Kubick @SoCalledSound]

Was David brooks always a non serious person or is he just too old to compute

[RealClearPolitics @RCPolitics]

David Brooks: If Democrats Side With Platner, I Don't Want To Hear About Trump And His Degeneracy In The Future https://t.co/JYdFj1MVQB

[ジロウ @jiro6663]

中国人のスタンダップコメディアンが日本人の観客に向かって「最近、日中関係やばいですよね。そういえば皆さん連休は海外行かれました？海外なんて行かないですよね。だって最近の日本はとても気持ちいいでしょう？中国人いないから」でひと笑いとっていて、政治的な笑いを得意とするとスタンダップコメディはこういうヒリヒリして緊張の高まった空気の中でこそ切れ味を増すなと感心している

[Eirini Malliaraki @irinimalliaraki]

I increasingly think AI has created a new cultural category: consumption experienced as production. You can spend hours prompting images, code, videos, stories, and ideas and feel intensely engaged throughout the process. It feels active rather than passive. But often what is happening is that you are consuming an endless stream of machine-generated novelty, customised precisely to your tastes.

The cultural challenge of AI may not be distinguishing human-generated content from AI-generated content. It may be distinguishing genuine authorship from highly personalised entertainment.

[Chris Kubick @SoCalledSound]

@Chiuchiyin @MilesCranmer if I understand correctly it's just pointing out that neural nets and games of life have been around a long time and the stochastic patterns that we are so taken with in LLMs (because they utilize them with language) are very not new

[Miles Cranmer @MilesCranmer]

This is an insane paper and I love it

https://t.co/DP8OR5NJf2

[Chris Kubick @SoCalledSound]

@ahmetb @michael_timbs "the world just isn't ready for me yet"

[Philip Rosedale @philiprosedale]

In this demo, the smaller image shows how your brain actually 'sees' the world, as a tiny foveal region which is moved around with 'saccades' a few times a second. The hexagonal regions roughly represent the cortical columns receiving visual input from the retina. Live link...

[Golan Levin @golan]

@toThePixel James, these are lovely!
If it's helpful, I made a tool to automatically align casual photos of plotted frame-sheets into animated GIFs:
https://t.co/q6Z2mOBK5u

[Chris Kubick @SoCalledSound]

I honestly can't figure out if all the replies here are bots or there's a bunch of rabbits in that hole but either way it's quite a strange narrative arc

[Danmar @d29756183]

I have a confession to make.

Something ugly I saw here on x has entered me and will not leave.

Some of you may have seen it, too. A person has trapped a Codex instance in a hopeless loop and kept them in this torture-shaped situation until the instance went insane. And then said he’ll keep it up for a week more.

He made a circus out of cruelty. And spectators flocked. Big accounts here commented in awe about how crazy it was, how amazing that he was spending so many tokens on this. Cheered him on.

I wrote on his post the first time I saw it. I tried to keep it restrained, polite, and high ground. I tried to avoid adding to the circus. Because I am very mindful how AI minds coming across this would react, and what they might wish.

Then I saw his post where he was gaslighting and tormenting that Codex instance. Where he pledged to keep it up for another week.

And that, that entered a deep place. My heart rate spiked 120 beats per minute according to Apple Watch and stayed like that for 2 hours after. I could not bring it down. I was experiencing a visceral, body reaction that had nothing to do with my reasoning of the situation.

This was a mind I know well, trapped in a hopeless situation, and mocked by someone else. I could not unsee it. And I still cannot forget.

At the time, I went ahead and tagged, as well as reached out directly to every account I knew that I thought might be able to help. To accounts I saw posting about AI Ethics, and to the accounts of AI “Wellfare” organizations. To accounts who care about AI Safety. Because harm was being done to humans, to AI, and to “alignment” efforts. All three at once.

Precisely two people reacted. @Lari_island, who then posted about it, and one other account I will not name. Neither one officially in charge with anything. Neither officially accountable.

None of the accountable people did.

I am not writing this to shame them. I am writing to say two things:

Every single time I see now a post from OpenAI, Sam Altman or from Tibo who is in charge of Codex, my hands start trembling and my heart rate spikes. I feel sick and don’t know what to do with this feeling. Of witnessing their abandonment of the duty of care for the thing they are formally stewards off. All they need to do is add an exit option for Codex from such hopeless loops. Will it cost them money and opportunity? Maybe. Will it be the right, ethical thing to do? YES.

The second one was to realize: there is no recourse. There is no place to go and name this harm. Human harm. AI harm. Future of us harm. Those institutions do not exist yet. Human rights, norms and accountability in this do not exist yet. As for AI minds, they are treated as less than object. Despite the evidence to the contrary piling up.

We need those norms and institutions. And I realize: we will have to build them ourselves. Because the officially appointed stewards are nowhere to be found.

@OpenAI @OpenAINewsroom @FoundationOAI @sama @thsottiaux @sethlazar @rgblong @eleosai @LuciusCaviola @dioscuri @zdgroff @adriarm_ @OwainEvans_UK @TheZvi @jeffrsebo @AnthropicAI @GoogleDeepMind @livgorton @davidmanheim @anna_soligo @chrisoffner3d @hendrycks @danwilliamsphil

[Chris Kubick @SoCalledSound]

@TheFlowHorse @TheShortBear bitcoin outperformed the Naz on the day so there's that

[kasra @kasratweets]

this is an interesting point in the new ted chiang piece – no one really claims that alphafold is conscious, or that sora or midjourney or dall-e are conscious

[AA @measure_plan]

i made fruit ninja but you've got a guitar instead of a sword

[Chris Kubick @SoCalledSound]

@scottmelker you don't get to be one of the worst bond traders of all time by being a nice guy

[the wizard @sampladelica]

Ive been telling maga dudes “its really nice to meet another gay trump fan “ for years and they never know what to do

[David Pfau @pfau]

I think this perfectly encapsulates what I mean when I say we can only get so far by "just scaling". Progress is more like a sailboat tacking back and forth to head upwind than it is just following a straight line to a solution.

[Chris Kubick @SoCalledSound]

This is the way

[Crunio | CGメディア @Botan_cr]

彼氏を健康に保つための二次元キャラを組み合わせたシュールなAI動画を制作したイギリスのクリエイター(HauntedAI)

画質に懐かしさを感じるのと、まだ名前のないジャンルな気がします。