λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y)))

𝜆 f. (𝜆 x. f (x x)) (𝜆 x. f (x x))

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@haruhiprophet still genuinely insane to me that it ran on 7th gen at all

[pan 🇰🇮 @jgveq]

The left created this little guy to insult us.
We adopted him.

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

fork in kitchen etc etc

[Tom Siwik @tomhacks]

Omarchy folks. Less slop, more code reviews, please.

v3 -> v4 migration just deleted my network setup without warning. SSH gone, tailscale gone. And migration never finished bc of a browser bug.

That's not called a migration

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@airshaped you have to lead into it by being like there's NO way this could ever happen, right

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@airshaped this is killing me

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@kernel_trick hi

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

tweet about ambiguity recreates exact ambiguity it wanted to comment on

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@airshaped i even tried giving it data!

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@airshaped i tried the exact same and mine just produced a really uninteresting image of genres of music listed on a canvas

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@respirebtw I post abstract thoughts as if nobody will read them and then people read them and misinterpret me entirely </3

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

Can these NO AI 🚫 people stop interacting with my tweet

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@Tomycj1 https://t.co/5rbxvfjrPT

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

ok before this leaves my orbit any further:
i am not trying to say that llms aren't being used in bio nor am i trying to make a moral claim about llms or ai models more broadly.

the original poster wasn't saying that ai = good bc cancer they were saying that this is some narrowly defensible usage of ai; i am saying that there isn't such a binary between the two.

alphafold 3 for example is a "generative ai model", likewise gpt-rosalind and claude science etc are actually being used in the natural sciences and medicine.

"the ai that detects cancer early" (presumably smth like CHIEF) is a hybrid cnn-transformer visual model used for discriminatory tasks, so in fairness is not a generative model, but such generative models are broadly used in biomedical research

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

my point more accurately phrased, because i didn't expect this tweet to be seen by more than like 5 people was a plain observation that there is an interesting folk ontology of "good ai" (finds cancer) vs "bad ai" (makes games bad and ugly art)

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

ok before this leaves my orbit any further:
i am not trying to say that llms aren't being used in bio nor am i trying to make a moral claim about llms or ai models more broadly.

the original poster wasn't saying that ai = good bc cancer they were saying that this is some narrowly defensible usage of ai; i am saying that there isn't such a binary between the two.

alphafold 3 for example is a "generative ai model", likewise gpt-rosalind and claude science etc are actually being used in the natural sciences and medicine.

"the ai that detects cancer early" (presumably smth like CHIEF) is a hybrid cnn-transformer visual model used for discriminatory tasks, so in fairness is not a generative model, but such generative models are broadly used in biomedical research

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@philzona wasn't really what i meant, look at the context of the original tweet

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@f4micom what kinda throughput?

[うさぎのもか @moca_usa_mofu]

画面越しにたくさんの人に見られてるとも知らず、今日も一生懸命お顔を洗ってるうさぎ。

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@RoseSilicon i think you misinterpreted what i was saying; i meant more so that non-technical types seem to think there's some binary distinction between "cancer curing ai" and "bad ai" and that the two are cleanly separable

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@_0x0000000 are you disagreeing with me? bc i meant like, they have this distinct mental model of there being "good ai" ("vibe coding", image generation, etc.) and "bad ai", where there's some specific singular ai model that is used in bio and the other models are Evil and Theft

[λl₁ i l₂ l₃ y. l₁(i(l₂(l₃ y))) @bunniemoder]

@airshaped Mine is doing quite well for nearly a year old