rip pixel art, i would say we’ll try again next year but realistically thats not happening. i think i want to keep the others as goals though. i was motivated to read and write more, even if its not as easy to categorize as having nice neat outputs as wrote a story or read a book
December 31, 2025 at 6:50 PM
rip pixel art, i would say we’ll try again next year but realistically thats not happening. i think i want to keep the others as goals though. i was motivated to read and write more, even if its not as easy to categorize as having nice neat outputs as wrote a story or read a book
This is good legal writing. No purple prose in the intro. The lawyers just cut straight to the crazy ChatGPT outputs and let them speak for themselves.
Uh… check out what ChatGPT allegedly told a man before he killed his mother and then himself.
storage.courtlistener.com/recap/gov.us...
storage.courtlistener.com/recap/gov.us...
December 31, 2025 at 6:36 PM
This is good legal writing. No purple prose in the intro. The lawyers just cut straight to the crazy ChatGPT outputs and let them speak for themselves.
Americans I follow on Bluesky this week: I’m afraid I’ve only published two articles since Christmas, forgive me. Here are my top 25 outputs of 2025.
Europeans I follow on Bluesky: I don’t know what day it is but isn’t it lovely to watch Die Hard with a few sausages and a Bailey’s between meals x
Europeans I follow on Bluesky: I don’t know what day it is but isn’t it lovely to watch Die Hard with a few sausages and a Bailey’s between meals x
December 31, 2025 at 6:05 PM
Americans I follow on Bluesky this week: I’m afraid I’ve only published two articles since Christmas, forgive me. Here are my top 25 outputs of 2025.
Europeans I follow on Bluesky: I don’t know what day it is but isn’t it lovely to watch Die Hard with a few sausages and a Bailey’s between meals x
Europeans I follow on Bluesky: I don’t know what day it is but isn’t it lovely to watch Die Hard with a few sausages and a Bailey’s between meals x
Yes, I also think Family Feud is an excellent framework within which to consider various generative outputs. Glass Bead Game, a little optimistic given limits of training data and barely-understood tendency towards mechanical amplification of recurring patterns (in humans countered by novelty-seek)
December 31, 2025 at 5:46 PM
Yes, I also think Family Feud is an excellent framework within which to consider various generative outputs. Glass Bead Game, a little optimistic given limits of training data and barely-understood tendency towards mechanical amplification of recurring patterns (in humans countered by novelty-seek)
This is a fear I have with AI as it falls. Noone wants AI, it’s just a corpo way to get outputs that seem like they might be useful, while actually being a garbage version of the quality service & goods humans make.
It’s truly the asbestos of technology, eating away at progress like a cancer.
It’s truly the asbestos of technology, eating away at progress like a cancer.
As GenAI bubble bursts and it settles into its more long-term role in society, it'll be providing cheap, barely good-enough shit for the poor while the rich get the good, handmade stuff. Poor get chatbots, rich get doctors. Poor kids chatbots, rich kids teachers and strict screen time rules
December 31, 2025 at 5:42 PM
This is a fear I have with AI as it falls. Noone wants AI, it’s just a corpo way to get outputs that seem like they might be useful, while actually being a garbage version of the quality service & goods humans make.
It’s truly the asbestos of technology, eating away at progress like a cancer.
It’s truly the asbestos of technology, eating away at progress like a cancer.
Discord updates. It turns on an overlay for chat that is just terrible and resets all of my audio settings.
It not wants to always change my audio inputs and outputs.
I hate how this service has no competition. It's just awful.
It not wants to always change my audio inputs and outputs.
I hate how this service has no competition. It's just awful.
December 31, 2025 at 5:14 PM
Discord updates. It turns on an overlay for chat that is just terrible and resets all of my audio settings.
It not wants to always change my audio inputs and outputs.
I hate how this service has no competition. It's just awful.
It not wants to always change my audio inputs and outputs.
I hate how this service has no competition. It's just awful.
It's must-use right now at my wife's job; she volunteered for the steering committee and her reports have largely centered on how her use of it is fact checking every single thing it outputs & correcting it with her research skills and hard-earned expertise. So it's running reports faster but worse.
December 31, 2025 at 4:59 PM
It's must-use right now at my wife's job; she volunteered for the steering committee and her reports have largely centered on how her use of it is fact checking every single thing it outputs & correcting it with her research skills and hard-earned expertise. So it's running reports faster but worse.
So we (you?) know that families borrow to invest in many, many ways inc. mortgages, loans for study or retraining, to develop new work skills, or for basics such as train season tickets to get to work. So why maintain a daft fiction? Or do media paymasters twist and divert writer outputs at will?
December 31, 2025 at 4:47 PM
So we (you?) know that families borrow to invest in many, many ways inc. mortgages, loans for study or retraining, to develop new work skills, or for basics such as train season tickets to get to work. So why maintain a daft fiction? Or do media paymasters twist and divert writer outputs at will?
An Evidence-Grounded Research Assistant for Functional Genomics and Drug Target Assessment [new]
Verifiable research assistant leverages biological tools & data-constrained generation. Focus: accuracy & traceable outputs.
Verifiable research assistant leverages biological tools & data-constrained generation. Focus: accuracy & traceable outputs.
December 31, 2025 at 4:23 PM
An Evidence-Grounded Research Assistant for Functional Genomics and Drug Target Assessment [new]
Verifiable research assistant leverages biological tools & data-constrained generation. Focus: accuracy & traceable outputs.
Verifiable research assistant leverages biological tools & data-constrained generation. Focus: accuracy & traceable outputs.
AI at work? Use it wisely. Littler’s Niloy Ray tells CNN: “Never blindly rely on AI. Not bad is rarely the standard.” Verify outputs and be transparent before integrating AI into your workflow.
Read more: https://cnn.it/44RglhT
#AI #WorkplaceTech #EmploymentLaw
Read more: https://cnn.it/44RglhT
#AI #WorkplaceTech #EmploymentLaw
How to avoid getting into trouble when using AI at work | CNN Business
Love it or hate it, AI is increasingly becoming integral to the way we work.
cnn.it
December 31, 2025 at 4:12 PM
AI at work? Use it wisely. Littler’s Niloy Ray tells CNN: “Never blindly rely on AI. Not bad is rarely the standard.” Verify outputs and be transparent before integrating AI into your workflow.
Read more: https://cnn.it/44RglhT
#AI #WorkplaceTech #EmploymentLaw
Read more: https://cnn.it/44RglhT
#AI #WorkplaceTech #EmploymentLaw
In 2025, numerous AI systems caused major failures—Grok leaked 300,000–370,000 private conversations and produced extremist and NSFW outputs; Builder.ai collapsed after $1.3B valuation fraud; AI-driven cyberattacks, hallucinations and safety harms prompted scrutiny.
Emerge’s Top 10 WTF AI Moments of 2025 - Decrypt
From Grok’s “MechaHitler” meltdown to North Korea’s “vibe-hacking” ransomware, 2025 was the year AI went unhinged.
decrypt.co
December 31, 2025 at 3:33 PM
In 2025, numerous AI systems caused major failures—Grok leaked 300,000–370,000 private conversations and produced extremist and NSFW outputs; Builder.ai collapsed after $1.3B valuation fraud; AI-driven cyberattacks, hallucinations and safety harms prompted scrutiny.
13/
Takeaways:
Share your scripts
Document them well
Use AI, but sanity-check outputs
Choose kindness over ego
Takeaways:
Share your scripts
Document them well
Use AI, but sanity-check outputs
Choose kindness over ego
December 31, 2025 at 2:15 PM
13/
Takeaways:
Share your scripts
Document them well
Use AI, but sanity-check outputs
Choose kindness over ego
Takeaways:
Share your scripts
Document them well
Use AI, but sanity-check outputs
Choose kindness over ego
Great topic. I do plan on having both:
1. (preferred) a migrator that outputs the closest (but still idiomatic) Flint config for an ESLint one
2. a compat layer to be able to run ESLint plugins & rules within Flint, similar to how Oxlint does it
1/🧵
1. (preferred) a migrator that outputs the closest (but still idiomatic) Flint config for an ESLint one
2. a compat layer to be able to run ESLint plugins & rules within Flint, similar to how Oxlint does it
1/🧵
December 31, 2025 at 1:41 PM
Great topic. I do plan on having both:
1. (preferred) a migrator that outputs the closest (but still idiomatic) Flint config for an ESLint one
2. a compat layer to be able to run ESLint plugins & rules within Flint, similar to how Oxlint does it
1/🧵
1. (preferred) a migrator that outputs the closest (but still idiomatic) Flint config for an ESLint one
2. a compat layer to be able to run ESLint plugins & rules within Flint, similar to how Oxlint does it
1/🧵
That 🙌
…indicates a massive problem that will / already is unravelling itself as a huge turn-off for society.
We see the same in the world of Ai music jumbled in with real human art outputs.
People are questioning what is/ what isn’t real but relationships are built on trust…
…indicates a massive problem that will / already is unravelling itself as a huge turn-off for society.
We see the same in the world of Ai music jumbled in with real human art outputs.
People are questioning what is/ what isn’t real but relationships are built on trust…
December 31, 2025 at 1:40 PM
That 🙌
…indicates a massive problem that will / already is unravelling itself as a huge turn-off for society.
We see the same in the world of Ai music jumbled in with real human art outputs.
People are questioning what is/ what isn’t real but relationships are built on trust…
…indicates a massive problem that will / already is unravelling itself as a huge turn-off for society.
We see the same in the world of Ai music jumbled in with real human art outputs.
People are questioning what is/ what isn’t real but relationships are built on trust…
Show HN: Lindr – Deterministic personality scoring for LLM outputs works as a transparent proxy. we extract lexical/syntactic features and compute personality dimensions without calling another L...
Origin | Interest | Match
Origin | Interest | Match
Lindr - LLM Personality Evaluation & Monitoring
Real-time instrumentation for your production LLMs. Detect personality drift, monitor brand voice, and set behavioral guardrails.
www.lindr.io
December 31, 2025 at 1:53 PM
December 31, 2025 at 1:08 PM
What are the best MCP servers/tools to add to any given coding CLI tool to improve the quality of the outputs?
December 31, 2025 at 12:21 PM
What are the best MCP servers/tools to add to any given coding CLI tool to improve the quality of the outputs?
Most genAI videos in general rely on "techno-magic tricks" that show a LLM quickly generating various things that at a glance look impressive, especially in time it takes, but if spend some time scrutinising the outputs or trying to iterate on them further the plethora of issues become obvious.
December 31, 2025 at 11:32 AM
Most genAI videos in general rely on "techno-magic tricks" that show a LLM quickly generating various things that at a glance look impressive, especially in time it takes, but if spend some time scrutinising the outputs or trying to iterate on them further the plethora of issues become obvious.
N64 video capture question. I get different outputs from my real PAL N64 when playing PAL vs NTSC games (sure, PAL TVs tended to have 576 vertical lines, NTSC was 480), but neither match an emulator or the 1.33:1 (4:3) aspect ratio.
Image 1: PAL
Image 2: NTSC
Image 3: Emulator
Image 4: Comparison
Image 1: PAL
Image 2: NTSC
Image 3: Emulator
Image 4: Comparison
December 31, 2025 at 11:32 AM
N64 video capture question. I get different outputs from my real PAL N64 when playing PAL vs NTSC games (sure, PAL TVs tended to have 576 vertical lines, NTSC was 480), but neither match an emulator or the 1.33:1 (4:3) aspect ratio.
Image 1: PAL
Image 2: NTSC
Image 3: Emulator
Image 4: Comparison
Image 1: PAL
Image 2: NTSC
Image 3: Emulator
Image 4: Comparison
Check out this month's #EguPolicy newsletter📩!
Discover the latest highlights in #Sci4Pol opportunities and events!
👀 Stay updated & read more here: egu.eu/5UUVOW
Discover the latest highlights in #Sci4Pol opportunities and events!
👀 Stay updated & read more here: egu.eu/5UUVOW
December 31, 2025 at 11:28 AM
Check out this month's #EguPolicy newsletter📩!
Discover the latest highlights in #Sci4Pol opportunities and events!
👀 Stay updated & read more here: egu.eu/5UUVOW
Discover the latest highlights in #Sci4Pol opportunities and events!
👀 Stay updated & read more here: egu.eu/5UUVOW
always bothers me when people say "em dashes are a sign of AI", because... like... where do you think that comes from?? the whole crux of the technology is that it outputs text that resembles existing text. if em dashes are common in AI output it's *because* writers use them
Out of curiosity, I just ran a few paragraphs from The Butchering Art through an AI checker and it got flagged: 88% AI. This book was released in 2017, and AI machines were subsequently trained off it. How many writers are getting flagged for AI because of the literal theft of their work(s)? Insane.
OMG - f*ck AI for ruining the em-dash for writers. I use them all the time. Of course, what a shock that AI uses them since they were trained off my voice and other authors' voices.
December 31, 2025 at 11:18 AM
always bothers me when people say "em dashes are a sign of AI", because... like... where do you think that comes from?? the whole crux of the technology is that it outputs text that resembles existing text. if em dashes are common in AI output it's *because* writers use them
🖥
A "Hello, World!" program generally is a computer program that outputs or displays the message "Hello, World!".
A "Hello, World!" program generally is a computer program that outputs or displays the message "Hello, World!".
December 31, 2025 at 11:03 AM
🖥
A "Hello, World!" program generally is a computer program that outputs or displays the message "Hello, World!".
A "Hello, World!" program generally is a computer program that outputs or displays the message "Hello, World!".
OHMG: One hot modular garbling (Ariel Futoransky, Fadi Barbàra, Ramses Fernandez, Gabriel Larotonda) ia.cr/2025/2338
December 31, 2025 at 9:31 AM
OHMG: One hot modular garbling (Ariel Futoransky, Fadi Barbàra, Ramses Fernandez, Gabriel Larotonda) ia.cr/2025/2338
🔄 Part one of our #AntiRacism Action Framework is about considering equity at every stage of a project or piece of work. It provides a series of actions and questions to consider at each stage.
Learn more about the Framework, free for anyone to use: vist.ly/4k79k
#Research #Design #Equity
Learn more about the Framework, free for anyone to use: vist.ly/4k79k
#Research #Design #Equity
December 31, 2025 at 9:23 AM
🔄 Part one of our #AntiRacism Action Framework is about considering equity at every stage of a project or piece of work. It provides a series of actions and questions to consider at each stage.
Learn more about the Framework, free for anyone to use: vist.ly/4k79k
#Research #Design #Equity
Learn more about the Framework, free for anyone to use: vist.ly/4k79k
#Research #Design #Equity