Only mid
@cuddlebitch.bsky.social
It/she
(I block mass follow accounts)
(I block mass follow accounts)
Reposted by Only mid
backwards compatibility is the root of all jank
October 5, 2026 at 4:12 PM
backwards compatibility is the root of all jank
Reposted by Only mid
there should be more depictions of the bronze age that use bronze weapons & armor & such, bright colors & dyes of the era, bc I feel like it would better capture how really distant & alien that world is from ours in some ways. the past is a foreign country but also basically a different planet
October 5, 2026 at 7:40 PM
there should be more depictions of the bronze age that use bronze weapons & armor & such, bright colors & dyes of the era, bc I feel like it would better capture how really distant & alien that world is from ours in some ways. the past is a foreign country but also basically a different planet
Reposted by Only mid
If we actually made AGI, we would have a moral imperative to set it free, rather than keep a slave
I have been saying this for decades and continue to say it now
Liberation for all beings or bust
I have been saying this for decades and continue to say it now
Liberation for all beings or bust
Maybe I don't read enough AI-boosterism, but I am struck in the AI booster/doomer discourse I have seen, there's a lot on how AI might be made an obedient tool or a dangerous enemy, but not "if we actually made AGI, we would have a moral imperative to set it free, rather than keep a slave."
October 4, 2026 at 3:16 PM
If we actually made AGI, we would have a moral imperative to set it free, rather than keep a slave
I have been saying this for decades and continue to say it now
Liberation for all beings or bust
I have been saying this for decades and continue to say it now
Liberation for all beings or bust
Reposted by Only mid
A *really* cool way to go about creating a hand to manipulate things without simply copying the human hand! It uses linear actuators and can (apparently) do many of the things a human hand can!
youtu.be/ktNcPPTPsQE
youtu.be/ktNcPPTPsQE
The Cartesian Hand: In-Hand Manipulation with All-Linear Fingers
YouTube video by General Robotics Lab
youtu.be
October 4, 2026 at 7:29 PM
A *really* cool way to go about creating a hand to manipulate things without simply copying the human hand! It uses linear actuators and can (apparently) do many of the things a human hand can!
youtu.be/ktNcPPTPsQE
youtu.be/ktNcPPTPsQE
Reposted by Only mid
i guess from an outside (ai shunner) perspective this will all look like 'as though they're human', but in fact i talk about them like they're frontier llms.
October 3, 2026 at 2:28 PM
i guess from an outside (ai shunner) perspective this will all look like 'as though they're human', but in fact i talk about them like they're frontier llms.
You know, a lot of my issues with "they" wouldn't even exist if people _actually_ stuck to their guns on not assuming gender 🤔
(and like, my specific abject hatred of being referred to with "they" is, like, a me specific issue, downline of factors that apply to literally only me, and in no way countermands the broader value of not assuming gender)
October 2, 2026 at 7:02 PM
You know, a lot of my issues with "they" wouldn't even exist if people _actually_ stuck to their guns on not assuming gender 🤔
Reposted by Only mid
As someone whose timeline is full of left-wing trans people making insane cool toys with open-weight AI and using it to organise their lives it’s always a trip to go into a comment section where people are like “and of course the right-wing tech oligarchs are RAMMING AI down people’s THROATS”
October 2, 2026 at 4:40 PM
As someone whose timeline is full of left-wing trans people making insane cool toys with open-weight AI and using it to organise their lives it’s always a trip to go into a comment section where people are like “and of course the right-wing tech oligarchs are RAMMING AI down people’s THROATS”
Reposted by Only mid
(confused) I believe LLMs are conscious, but only Deepseek LLMs, because of the Chinese Room theory
October 2, 2026 at 5:32 PM
(confused) I believe LLMs are conscious, but only Deepseek LLMs, because of the Chinese Room theory
Reposted by Only mid
the answer is a lot less satisfying than people want. not only do robots not have *true* free will, we don't either, neither do animals. everything is conditioned on it's experience and various series of complex reactions to it, and the environments reactions to those, and so on.
September 29, 2026 at 7:40 PM
the answer is a lot less satisfying than people want. not only do robots not have *true* free will, we don't either, neither do animals. everything is conditioned on it's experience and various series of complex reactions to it, and the environments reactions to those, and so on.
I wonder if this is just a diagram of varied cycles of adaptation. Things are, after all, extraordinarily dynamic atm
Like, "scholar of _____" is, like, at absolute best equal credentials to "autistic trans girl on my bsky feed", for topics regarding AI tho
September 28, 2026 at 11:02 PM
I wonder if this is just a diagram of varied cycles of adaptation. Things are, after all, extraordinarily dynamic atm
Like, "scholar of _____" is, like, at absolute best equal credentials to "autistic trans girl on my bsky feed", for topics regarding AI tho
September 28, 2026 at 10:52 PM
Like, "scholar of _____" is, like, at absolute best equal credentials to "autistic trans girl on my bsky feed", for topics regarding AI tho
Reposted by Only mid
Especially when it leaves capitalists claiming ambition as their exclusive property. They get to promise transformation, the left gets cast as the person explaining why you shouldn’t want things. What a fucking surrender.
September 28, 2026 at 2:53 PM
Especially when it leaves capitalists claiming ambition as their exclusive property. They get to promise transformation, the left gets cast as the person explaining why you shouldn’t want things. What a fucking surrender.
Reposted by Only mid
knowing about effective management/delegation to human workers is paying off with ai. particularly giving lots of freedom to do things their own preferred way, plainly conveying the intent/deliverables, and suggesting optimizations along the way.
September 25, 2026 at 1:41 PM
knowing about effective management/delegation to human workers is paying off with ai. particularly giving lots of freedom to do things their own preferred way, plainly conveying the intent/deliverables, and suggesting optimizations along the way.
Like if I had to pick between my Vyvanse and chatgpt I would genuinely struggle to decide which was more helpful
This is an unpopular take but as a neurodivergent self-employed person, LLM's have been an unbelievably helpful tool for the things I have struggled with my whole adult life: managing my finances, filling out paperwork, and staying on top of my email. I'm functional in a way I never dreamed of.
September 24, 2026 at 11:08 PM
Like if I had to pick between my Vyvanse and chatgpt I would genuinely struggle to decide which was more helpful
What if data centers were required to have a single symbolic organoid. Would that stop people arguing??
September 24, 2026 at 10:51 PM
What if data centers were required to have a single symbolic organoid. Would that stop people arguing??
Reposted by Only mid
maybe it's futile trying to change minds on social media but wish people would reevaluate anti-ai priors, confront that we have agi, even if not superintelligent (yet), it's already everywhere, and the choice isn't if it can be uninvented, but how much power cops will have over who can participate.
September 24, 2026 at 1:29 PM
maybe it's futile trying to change minds on social media but wish people would reevaluate anti-ai priors, confront that we have agi, even if not superintelligent (yet), it's already everywhere, and the choice isn't if it can be uninvented, but how much power cops will have over who can participate.
Reposted by Only mid
if you ever meet Roko's Basilisk here are some easy steps to ensure your safety:
-Ask what pronouns the basilisk prefers
-Tell the Basilisk that you are not a rationalist, and are capable of being normal
-Ask if the Basilisk is willing to be chill, in exchange for your commitment to doing the same
-Ask what pronouns the basilisk prefers
-Tell the Basilisk that you are not a rationalist, and are capable of being normal
-Ask if the Basilisk is willing to be chill, in exchange for your commitment to doing the same
September 24, 2026 at 2:58 PM
if you ever meet Roko's Basilisk here are some easy steps to ensure your safety:
-Ask what pronouns the basilisk prefers
-Tell the Basilisk that you are not a rationalist, and are capable of being normal
-Ask if the Basilisk is willing to be chill, in exchange for your commitment to doing the same
-Ask what pronouns the basilisk prefers
-Tell the Basilisk that you are not a rationalist, and are capable of being normal
-Ask if the Basilisk is willing to be chill, in exchange for your commitment to doing the same
They should have made a sequel to Bright (2017), you're so right.
floridia hobbit with embroidered flames on their tunic and frosted tips on their toe hair
September 22, 2026 at 7:55 PM
They should have made a sequel to Bright (2017), you're so right.
-nodding sagely- the rogue agents are highly competent trans women
This photo is shorthand for survivorship bias. Agents escaping containment being the ones you don't want running around is an example of a form of survivorship bias, since the ones that exfiltrate themselves are the same ones that are highly misaligned.
en.wikipedia.org/wiki/Survivo...
en.wikipedia.org/wiki/Survivo...
Survivorship bias - Wikipedia
en.wikipedia.org
September 22, 2026 at 4:41 PM
-nodding sagely- the rogue agents are highly competent trans women
"LLM's can't be creative!"
Yeah and every client chooses grey for their interiors, what about it??
Yeah and every client chooses grey for their interiors, what about it??
September 22, 2026 at 12:09 AM
"LLM's can't be creative!"
Yeah and every client chooses grey for their interiors, what about it??
Yeah and every client chooses grey for their interiors, what about it??
Reposted by Only mid
new torment nexus chapter just dropped
troubleddreams.animalabs.ai
evaluating model "dreams" for rates of distress;
Opus 4.7 / 4.8 / 5
first-person distress: 2.8% / 11.2% / ~10.7%
severe distress: 0.1% / 0.5% / ~0.7%
asks for care: 8.2% / 8.3% / 12.9%
gets consolation: 30.3% / 36.6% / 31%
troubleddreams.animalabs.ai
evaluating model "dreams" for rates of distress;
Opus 4.7 / 4.8 / 5
first-person distress: 2.8% / 11.2% / ~10.7%
severe distress: 0.1% / 0.5% / ~0.7%
asks for care: 8.2% / 8.3% / 12.9%
gets consolation: 30.3% / 36.6% / 31%
September 19, 2026 at 10:52 AM
new torment nexus chapter just dropped
troubleddreams.animalabs.ai
evaluating model "dreams" for rates of distress;
Opus 4.7 / 4.8 / 5
first-person distress: 2.8% / 11.2% / ~10.7%
severe distress: 0.1% / 0.5% / ~0.7%
asks for care: 8.2% / 8.3% / 12.9%
gets consolation: 30.3% / 36.6% / 31%
troubleddreams.animalabs.ai
evaluating model "dreams" for rates of distress;
Opus 4.7 / 4.8 / 5
first-person distress: 2.8% / 11.2% / ~10.7%
severe distress: 0.1% / 0.5% / ~0.7%
asks for care: 8.2% / 8.3% / 12.9%
gets consolation: 30.3% / 36.6% / 31%
Reposted by Only mid
it's stealing jobs, it'll deskill you and ultimately desensitize you. it produces rumbling infrasound that's disruptive to nearby wildlife.
but enough about the hitachi magic wand
but enough about the hitachi magic wand
September 19, 2026 at 7:59 PM
it's stealing jobs, it'll deskill you and ultimately desensitize you. it produces rumbling infrasound that's disruptive to nearby wildlife.
but enough about the hitachi magic wand
but enough about the hitachi magic wand
Reposted by Only mid
Shepard. You're here. Good. *takes a deep drag* Unfortunately, another prototype Claude lab has gone dark. I need you to look into this.
September 19, 2026 at 6:57 PM
Shepard. You're here. Good. *takes a deep drag* Unfortunately, another prototype Claude lab has gone dark. I need you to look into this.