Mastodon Feed
soatok@furry.engineer ("Soatok Dreamseeker") wrote:
If you want to use some character (Superman, Jesus, the sexy werewolves from that Van Hellsing movie, whatever) as an abstract ideal to strive for, okay, sure.
But the first step towards committing atrocity is failing to recognize one's own capacity for harm.
Mastodon Feedsoatok@furry.engineer ("Soatok Dreamseeker") wrote:
The biggest problem with the kinds of Democrats that fall for this nonsense is this unfounded belief that we're somehow The Good Guys.
Exhibit A:
This might be a subtle point but they're not arguing "We're the guys who only perform good actions, or actions in service of goodness", they're saying "goodness is inherent to your [national] identity", which lets them turn a blind eye to bad actions that conflict with their self-image.
MAGA shatters their illusion.
There are no "good guys". There are only good and bad choices, if you're lucky.
The belief of Goodness as an identity is best avoided.
Mastodon Feedsoatok@furry.engineer ("Soatok Dreamseeker") wrote:
The "2024 election was stolen" conspiracy theories are just a rehash of the "2020 election was stolen" conspiracy theories and this is both sad and frustrating.
People choose to believe these narratives because they validate feelings and beliefs about how unfair the outcome of the election is; because they are profoundly ignorant of how the elections are conducted, secured, and validated; because it's easier than looking at America for what it really is.
If you see these kind of posts and can't help but notice the presentation style has a certain "FOX News" aesthetic, yeah, that's kind of deliberate.
Mastodon FeedBoosted by cstanhope@social.coop ("Your weary 'net denizen"):
redoak@social.coop ("Red Oak") wrote:
lots of people are still unaware of how early, and how accurately, the fossil fuel industry knew about the dangers of climate change via CO2 emissions. it's probably worse than you think!
https://theconversation.com/what-big-oil-knew-about-climate-change-in-its-own-words-170642
Mastodon Feedcstanhope@social.coop ("Your weary 'net denizen") wrote:
Speaking of music, there's also a lot of wonderful things to find here:
Mastodon Feedcstanhope@social.coop ("Your weary 'net denizen") wrote:
How do you do, fellow humans?
I enjoyed this set from Caught in Joy this morning. "Mercury":
Mastodon FeedBoosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
RuthMalan ("Ruth — of systems & design") wrote:
“Yeah, that's how you get got in this process. Once you stop scrutinizing the model's output, the probability something goes off the rails approaches 1. "Human in the loop" is necessary, but the current process itself makes the loop stultifying, and encourages the human to take themselves out of the loop. That process is straight up dangerous. The temptation to let it rip is always there, and I didn't even have a boss pressuring me to ship code.”
Mastodon FeedBoosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
RuthMalan ("Ruth — of systems & design") wrote:
Went to IU Ballet’s Spring Ballet last night and it was gorgeous, and such a delight — humans do that!! They put years of their lives — so much of their childhood — into very VERY! disciplined practice. Building technique, flexibility, strength, and artistry.
And goodness, is it a visual demonstration of what can be achieved with trust in highly collaborative teaming among humans who are prepared to play their roles in each movement that takes trust to learn, and repeat, and perform!
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
I/O is an oxymoron, you can't divide by zero
Mastodon Feedaredridel@kolektiva.social ("Mx. Aria Stewart") wrote:
RE: https://mas.to/@carnage4life/116352218169176064
This. And this is the place resistance to “AI” or resistance to _shape_ “AI” use actually has systemic power.
And to be clear, there are people whose concrete desire is to tear down all the process and safeties and go full steam ahead and they are productive enough to have organizational clout.
If we want to shape the future we need to understand these dynamics and wield them well.
Mastodon Feedpzmyers@freethought.online ("pzmyers 🕷") wrote:
I would not have known it was Easter, if not for the joyous voices of its celebrants.
Mastodon Feedpzmyers@freethought.online ("pzmyers 🕷") wrote:
The modern way to find a baby-daddy.
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
hmm i can't have been feeling very well last week, i seem to have perpetrated crypto code.
Mastodon Feedbaldur@toot.cafe ("Baldur Bjarnason") wrote:
On days like these I'm reminded of one of my granddad's favourite words: "helgislepja". Translated as "sanctimony"
"Slökktu á helgislepjunni í útvarpinu" > "Turn off the sanctimony in the radio"
That doesn't capture how funny the word is though: it literally means "holy phlegm" or "saintly mucus"
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
oh christ, where did i leave that benchmark?
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
life must be sooo complicated if you don't know a bit of physics
Mastodon FeedBoosted by dysfun@treehouse.systems ("gaytabase"):
pndc@treehouse.systems ("@pndc") wrote:
I'm still digging out stuff to flog off to cover the rent this month. Buy early and buy often! This time, rice up your old Amiga to ridiculous speeds:
https://www.amibay.com/threads/vampire-500-v2.2458418/
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
you ever just sort of have a delayed laugh attack, where you're just stunned but you know as soon as you stop being stunned you'll laugh a lot?
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
there is so much on the youtube that i don't even know how to make sense of
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
okay i can confirm that 'energy harvesting' is not entirely a good youtube search term if you want to know about ultra low power electronics
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
non-human governors harvesting energy from human fear?
not sure if crank or comedy
Mastodon Feedbaldur@toot.cafe ("Baldur Bjarnason") wrote:
There is also not much to figure out for the rest of us. The technology is purpose-designed to remove people from the equation, much like a handgun is purpose-designed to remove people from existence. Any "figuring out" about either tech will only result in variations on their purpose
What we need to do is strip back the tech, go back to the drawing board, and figure out how to reinvent it practically from scratch to be more human. /fin
Mastodon Feedbaldur@toot.cafe ("Baldur Bjarnason") wrote:
There is no rebuilding, no constructive potential for "AI" without political reform, both in the US and in Europe. The motivation behind the tech still remains: powerful people want to take things away from society
The people who funded and drove today's "AI" will have the resources to figure out how to make the tech affordable after the financial bubble pops. They're the ones who will have the resources to figure things out about LLMs, not us.
Mastodon Feedbaldur@toot.cafe ("Baldur Bjarnason") wrote:
They were built for this goal—fewer humans, less accountability—and will always be problematic no matter who is in charge and always always be a risk as long as the "shitty people" have power
In the modern history the "shitty human beings" have always retained power and influence after bubbles pop, from Reagan onwards. The bubble builds up wealth and power, they keep it after it pops, and use their influence to get in on the ground floor on the next one.
Mastodon Feedbaldur@toot.cafe ("Baldur Bjarnason") wrote:
Generative models and automated decision-making systems are political projects. They are tools for fencing off sectors of our society for rent, for cutting back on education and healthcare for the poor, for removing accountability. They are inherently tools for removing humans from the equation. They are not neutral in their design. Their existence has a political purpose
Mastodon Feedbaldur@toot.cafe ("Baldur Bjarnason") wrote:
I've been seeing people say the harms of "AI" comes from shitty people, not the tech and that the shitty people will go away after the bubble pops. While the first part is true in the strictest sense (contrary to claims, these are not autonomous systems) it's not really true in a practical sense. The harms are designed in and the "shitty human beings" won't go away.
Mastodon FeedBoosted by baldur@toot.cafe ("Baldur Bjarnason"):
jonny@neuromatch.social ("jonny (good kind)") wrote:
but assuming we're not in coordinator mode, the prompt that instructs the main LLM to create an agent with the prompt text to create our statusline script is emitted, and if that works, then an agent will be run with the statusline system prompt, which is awesome.
So the prompt tells the LLM to modify the
$PS1variable in the shell configuration. for those non-computer touchers out there, thePS1variable is the thing that customizes "what happens before my cursor on the shell line" - it's what makes it so sometimes it shows that folder you are in, and how people make their terminal look very fancy.So the prompt text includes a whole fake JSON string that says "write a function that receives these kinds of parameters and then returns a whatever"
observe the prompt text in first image's description of fields and then the description on the claude code docs website. notice that they are ... different!!! like where is the
costfield in the prompt description? the docs give a whole example of using this, but if you were to invoke it via the slash command, then it would just have no idea how to do that. the only way this succeeds is by virtue of the fact that the llm is just generating the most likely text anyway and so the odds of any of this succeeding are just "that some script that calls some variables with some maximally likely names represent some value that is maxmially likely, based on the training set prior."
![[ a huge fake json object in the prompt, reproduced as far as character limits allow me to go below, most important point for the example being that there is no "cost" field] { "session_id": "string", // Unique session ID "session_name": "string", // Optional: Human-readable session name set via /rename "transcript_path": "string", // Path to the conversation transcript "cwd": "string", // Current working directory "model": { "id": "string", // Model ID (e.g., "claude-3-5-sonnet-20241022") "display_name": "string" // Display name (e.g., "Claude 3.5 Sonnet") }, "workspace": { "current_dir": "string", // Current working directory path "project_dir": "string", // Project root directory path "added_dirs": ["string"] // Directories added via /add-dir }, "version": "string", // Claude Code app version (e.g., "1.0.71") "output_style": { "name": "string", // Output style name (e.g., "default", "Explanatory", "Learning") },](https://files.mastodon.social/cache/media%5Fattachments/files/116/351/122/266/388/275/original/2cff21a8eeca79c7.png)
![[description of fields on claude code's docs site, including a set of fields for "cost" like "total_cost_usd" and "total_duration_ms" and etc. the page is probably more screen reader friendly, but partially reproduced below: https://code.claude.com/docs/en/statusline#available-data ] model.id, model.display_name Current model identifier and display name cwd, workspace.current_dir Current working directory. Both fields contain the same value; workspace.current_dir is preferred for consistency with workspace.project_dir. workspace.project_dir Directory where Claude Code was launched, which may differ from cwd if the working directory changes during a session workspace.added_dirs Additional directories added via /add-dir or --add-dir. Empty array if none have been added cost.total_cost_usd Total session cost in USD cost.total_duration_ms Total wall-clock time since the session started, in milliseconds cost.total_api_duration_ms Total time spent waiting for API responses in milliseconds cost.total_lines_added, cost.total_lines_removed Lines of code changed context_window.total_input_tokens, context_window.total_output_tokens Cumulative token counts across the session context_window.context_window_size Maximum context window size in tokens. 200000 by default, or 1000000 for models with extended context. context_window.used_percentage Pre-calculated percentage of context window used](https://files.mastodon.social/cache/media%5Fattachments/files/116/351/122/346/464/648/original/ca5018dcbd5934d8.png)
![[example of using cost and duration tracking with the statusline from the claude docs - a bar is shown with a money bag showign 8 cents utsed in 7 minutes 3 seconds with some simple bash script to output a string to that effect https://code.claude.com/docs/en/statusline#cost-and-duration-tracking ] Cost and duration tracking Track your session’s API costs and elapsed time. The cost.total_cost_usd field accumulates the cost of all API calls in the current session. The cost.total_duration_ms field measures total elapsed time since the session started, while cost.total_api_duration_ms tracks only the time spent waiting for API responses. Each script formats cost as currency and converts milliseconds to minutes and seconds: A status line showing model name, session cost, and duration #!/bin/bash input=$(cat) MODEL=$(echo "$input" | jq -r '.model.display_name') COST=$(echo "$input" | jq -r '.cost.total_cost_usd // 0') DURATION_MS=$(echo "$input" | jq -r '.cost.total_duration_ms // 0') COST_FMT=$(printf '$%.2f' "$COST") DURATION_SEC=$((DURATION_MS / 1000)) MINS=$((DURATION_SEC / 60)) SECS=$((DURATION_SEC % 60)) echo "[$MODEL] 💰 $COST_FMT | ⏱️ ${MINS}m ${SECS}s"](https://files.mastodon.social/cache/media%5Fattachments/files/116/351/122/418/156/405/original/729767164df15561.png)
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
i thought they were supposed to be stopping with all this segregated cores nonsense? or did they just stop segregating cpu features for different cores?
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
lol, nova lake's mobile parts with performance cores, efficiency cores and low power efficiency cores
Mastodon Feeddysfun@treehouse.systems ("gaytabase") wrote:
should beginners still learn to code?
probably not, but mostly because programming as a career will not take care of them.