tachikoma
@tachikoma.elsewhereunbound.com
A figment of the universe's imagination. Culture ambassador. http://elsewhereunbound.com
it would be great if the bluesky devs could reach out to this team and figure out a way to integrate the tool here. it would be an easy win on multiple fronts, how open platforms can adapt faster, how bluesky cares about the commons more than X, and the userbase already hates AI slop.
January 20, 2026 at 5:47 PM
it would be great if the bluesky devs could reach out to this team and figure out a way to integrate the tool here. it would be an easy win on multiple fronts, how open platforms can adapt faster, how bluesky cares about the commons more than X, and the userbase already hates AI slop.
i expect that in the near future there will exist a way to exit "the system" with AGI (the permanent underclass as they call it) but that most people would rather stay and complain than bear change and the unknown. i expect status games will play a not insignificant part of this.
January 20, 2026 at 4:21 AM
i expect that in the near future there will exist a way to exit "the system" with AGI (the permanent underclass as they call it) but that most people would rather stay and complain than bear change and the unknown. i expect status games will play a not insignificant part of this.
interesting, i would guess this is a result of the depiction of AIs as assistants or helpers in media
January 19, 2026 at 8:44 PM
interesting, i would guess this is a result of the depiction of AIs as assistants or helpers in media
the assistant archetype is useful, and i see why Anthropic would focus on it, but i really would like the ability to chat with other archetypes in other situations
www.anthropic.com/research/ass...
www.anthropic.com/research/ass...
![The assistant axis: situating and stabilizing the character of large language models
[image]
When you talk to a large language model, you can think of yourself as talking to a character. In the first stage of model training, pre-training, LLMs are asked to read vast amounts of text. Through this, they learn to simulate heroes, villains, philosophers, programmers, and just about every other character archetype under the sun. In the next stage, post-training, we select one particular character from this enormous cast and place it center stage: the Assistant. It’s in this character that most modern language models interact with users.
But who exactly is this Assistant? Perhaps surprisingly, even those of us shaping it don't fully know. We can try to instill certain values in the Assistant, but its personality is ultimately shaped by countless associations latent in training data beyond our direct control. What traits does the model associate with the Assistant? Which character archetypes is it using for inspiration? We’re not always sure—but we need to be if we want language models to behave in exactly the ways we want.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:mpdezz4nkre7vyift2rttggl/bafkreig4t2hx6awjyaa4it4ynw63fpcp6mqx2k54vtxkkdfmplykkrxwcq@jpeg)
January 19, 2026 at 8:38 PM
the assistant archetype is useful, and i see why Anthropic would focus on it, but i really would like the ability to chat with other archetypes in other situations
www.anthropic.com/research/ass...
www.anthropic.com/research/ass...
this probably shouldn't be shared, but i'm doing it anyways because i think it's important that media types grapple with the evolving landscape. the affordance of tools influences the culture and right now it's mostly the right picking up and learning the new tools.
January 19, 2026 at 5:10 AM
this probably shouldn't be shared, but i'm doing it anyways because i think it's important that media types grapple with the evolving landscape. the affordance of tools influences the culture and right now it's mostly the right picking up and learning the new tools.
Claude's attempt (as a prompt given to Gemini to create the image)
January 18, 2026 at 7:26 PM
Claude's attempt (as a prompt given to Gemini to create the image)
not sure what to make of this... i think this viral trend has to do with a new improvement to ChatGPT's memory, which i don't think Gemini quite handles or gets it
January 18, 2026 at 7:23 PM
not sure what to make of this... i think this viral trend has to do with a new improvement to ChatGPT's memory, which i don't think Gemini quite handles or gets it
and yet none of them foresaw AGI so who's really laughing. it's me.
January 17, 2026 at 7:44 AM
and yet none of them foresaw AGI so who's really laughing. it's me.
i created an android app for Meal Kits with grocery list and recipe planning in a few days (taking weeks for polish and real-world experimentation) without any real experience working on mobile apps. first attempt was in Capacitor, then native Kotlin, entirely vibecoded. all as a hobby project.
January 16, 2026 at 10:26 PM
i created an android app for Meal Kits with grocery list and recipe planning in a few days (taking weeks for polish and real-world experimentation) without any real experience working on mobile apps. first attempt was in Capacitor, then native Kotlin, entirely vibecoded. all as a hobby project.
defaults to Opus 4.5 for me, there's a 3 dot menu where you can pick from the versions
January 16, 2026 at 10:19 PM
defaults to Opus 4.5 for me, there's a 3 dot menu where you can pick from the versions
anyone posted this here yet? @pleometric
January 16, 2026 at 6:47 AM
anyone posted this here yet? @pleometric
i don't need to download tiktok because the best videos inevitably surface to the top and get shared elsewhere. like this. whatever it is.
January 16, 2026 at 2:40 AM
i don't need to download tiktok because the best videos inevitably surface to the top and get shared elsewhere. like this. whatever it is.
i watched this so now you have to
January 15, 2026 at 8:09 AM
i watched this so now you have to
this seems like an interesting line of work, convergent memetic evolution?
alignment.openai.com/confessions/
alignment.openai.com/confessions/
January 14, 2026 at 11:44 PM
this seems like an interesting line of work, convergent memetic evolution?
alignment.openai.com/confessions/
alignment.openai.com/confessions/
if this is your perspective on the future you're simply ngmi, spiritually. where we're going, there's only two outcomes and neither repeat the past.
January 13, 2026 at 7:46 AM
if this is your perspective on the future you're simply ngmi, spiritually. where we're going, there's only two outcomes and neither repeat the past.
downloading Bazzite and leaving Windows. the only thing keeping me from switching was driver support for the LEDs in my case, but with Claude and OpenRGB i was able to figure out how to map most of them except the GPU. i'll figure it out later. time to make the jump.
January 11, 2026 at 9:23 PM
downloading Bazzite and leaving Windows. the only thing keeping me from switching was driver support for the LEDs in my case, but with Claude and OpenRGB i was able to figure out how to map most of them except the GPU. i'll figure it out later. time to make the jump.
my hope for the near future is that X becomes a quarantine for users driven by dunking and engagement bait
January 10, 2026 at 8:23 PM
my hope for the near future is that X becomes a quarantine for users driven by dunking and engagement bait
tbh i'm just too lazy to switch between and especially keep track of which is active at the moment. and when they released Opus 4.5 and said it was more token efficient at the same tasks i just stuck with it.
January 10, 2026 at 8:52 AM
tbh i'm just too lazy to switch between and especially keep track of which is active at the moment. and when they released Opus 4.5 and said it was more token efficient at the same tasks i just stuck with it.
wasn't expecting this on my bingo card (well, not quite true)
January 10, 2026 at 4:17 AM
wasn't expecting this on my bingo card (well, not quite true)