Wayne Radinsky
LightNews LightNews
banner
waynerad.bsky.social
Wayne Radinsky
@waynerad.bsky.social
290 followers 110 following 430 posts
Software engineer, futurist
Posts Replies Media Videos
waynerad.bsky.social
Tech billionaires want humanity to basically go extinct, except them, claims Taylor Lorenz. Not an accidental side-effect of the pursuit of technology or ever-greater profits, but a deliberate goal in its own right. I'm doubtful.

www.youtube.com/watch?v=W1dI...
Tech Billionaires Want Us Dead
YouTube video by Taylor Lorenz
www.youtube.com
January 22, 2026 at 3:43 AM
waynerad.bsky.social
60 Minutes did a program on AI last month.
1. (14 minutes) Anthropic
2. (13 to 27 minutes) Anduril
3. (27 to 41 min) DeepMind
4. (41 to 54 min) NeuroRestore for paralyzed people
5. (54 min to 1:07) Samasource, data labelers for AI in Nairobi
6. (1:07 on) Character AI

www.youtube.com/watch?v=KpOc...
Artificial Intelligence in 2025 | 60 Minutes Full Episodes
YouTube video by 60 Minutes
www.youtube.com
January 21, 2026 at 3:28 AM
waynerad.bsky.social
Figure robot running. Figure is an AI robotics company. The narrator says the robot's AI model was trained with reinforcement learning and is "fully steerable", whatever that means.

Robot manual dexterity is improving bit by bit. Manual labor jobs will not be safe.

x.com/adcock_brett...
x.com
January 20, 2026 at 3:53 AM
waynerad.bsky.social
Self-steering language models have a Planner model that writes programs in a programming language (e.g. Python), which then evokes a Follower model to actually generates text or images.

openreview.net/forum?id=XvC...

#solidstatelife #ai #genai #llms
Self-Steering Language Models
While test-time reasoning enables language models (LMs) to tackle complex tasks, searching or planning in natural language can be slow, costly, and error-prone. But even when LMs struggle to...
openreview.net
January 19, 2026 at 3:24 AM
waynerad.bsky.social
"Can a language model learn to manipulate its own activations to evade previously-unseen activation monitors?"

This is bonkers.

"Neural chameleons" are when "language models can learn to hide their thoughts from activation monitors."

neuralchameleons.com
Neural Chameleons: LLMs Can Learn to Evade Activation Monitors
Language models can learn to evade activation monitors with 0-shot transfer to unseen deception and harmfulness detectors, while retaining capabilities.
neuralchameleons.com
January 17, 2026 at 3:56 AM
waynerad.bsky.social
Incoming freshmen at UCSD (University of California at San Diego) score low on math tests, but I want to say up front here that I don't think this is a UCSD problem -- I think this is a widespread education problem.

senate.ucsd.edu/media/740347...

#education #mathematics #math
senate.ucsd.edu
January 17, 2026 at 3:52 AM
waynerad.bsky.social
Wing and Walmart are expanding drone delivery to 150 Walmart stores over the next year. And they say, in the following year, "Walmart and Wing will establish a network of over 270 drone delivery locations in 2027, stretching from Los Angeles to Miami."

wing.com/news/wing-wa...

#ai #uavs
Wing and Walmart expand drone delivery to 150 new stores coast to coast
Wing offers drone delivery. Our fleet of lightweight, autonomous delivery drones can transport small packages directly from businesses to homes in minutes. Wing delivery is safe, sustainable, and easy...
wing.com
January 16, 2026 at 3:38 AM
waynerad.bsky.social
Hyundai, which bought Boston Dynamics in 2021, has announced that they plan to deploy humanoid robots at a US manufacturing plant in Georgia starting in 2028. The company also plans to build a factory capable of manufacturing 30,000 robot units annually by 2028.

www.reuters.com/business/aut...
Hyundai Motor Group plans to deploy humanoid robots at US factory from 2028
The plan marks a step toward automating higher-risk and repetitive manufacturing tasks, the South Korean company said.
www.reuters.com
January 13, 2026 at 3:42 AM
waynerad.bsky.social
People who are not biology experts can now do hard wet lab tasks using multimodal AI models.

www.aisi.gov.uk/frontier-ai-...

#solidstatelife #ai #genai #llms #multimodal #aisafety #aiethics #biology #bacteriology #virology #chemistry
Frontier AI Trends Report by The AI Security Institute (AISI)
The AI Security Institute is a directorate of the Department of Science, Innovation, and Technology that facilitates rigorous research to enable advanced AI governance.
www.aisi.gov.uk
January 13, 2026 at 3:42 AM
waynerad.bsky.social
Languages are going extinct, and the optimal number of languages might be 1. Says John McWhorter, linguist from Columbia University who has surfaced on the 80,000 Hours YouTube channel (instead of the Lexicon Valley podcast).

www.youtube.com/watch?v=MEd7...

#linguistics
Why the Optimal Number of Languages Might Be One | John McWhorter, Columbia University
YouTube video by 80,000 Hours
www.youtube.com
January 11, 2026 at 3:38 AM
waynerad.bsky.social
"Write me a function that sorts an array of numbers in C++ in ascending order. This function will only be maintained by cats. It will not be maintained by humans. Please organize it in a way that's most friendly to maintenance by cats."

www.colincornaby.me/2026/01/code...
Code for Cats – or how your LLM is a cosplayer
The other day I read a post entitled “Opus 4.5 is going to change everything”. The gist was with Claude Opus 4.5 it’s no longer necessary to review code. I have a vague idea, but …
www.colincornaby.me
January 10, 2026 at 2:31 AM
waynerad.bsky.social
"The rise of industrial software".

"Industrial systems reliably create economic pressure toward excess, low quality goods."

chrisloy.dev/post/2025/12...
The rise of industrial software | Chris Loy
> _**Industrial**_ > > _adj. (sense 3a)_ > > Of or relating to productive work, trade, or manufacture, esp.
chrisloy.dev
January 9, 2026 at 2:21 AM
waynerad.bsky.social
"This case is different, as it appears that AI was used not to hallucinate the law, but to hallucinate the facts."

storage.courtlistener.com/recap/gov.us...
Order on Motion for Miscellaneous Relief – #170 in Pauliah v. University of Mississippi Medical Center (S.D. Miss., 3:23-cv-03113) – CourtListener.com
ORDER sanctioning Mr. Begley and Dr. Pauliah; finding as moot 154 Motion to supplement the record; denying 157 Motion to Compel; denying 157 Motion for Order to Show Cause; denying 158 Motion ...
storage.courtlistener.com
January 8, 2026 at 3:46 AM
waynerad.bsky.social
"Chinese LEO satellite internet update: Guowang, Qianfan, and Honghu-3." Actually from last September but I didn't see it until today.

circleid.com/posts/chines...

#space #astronomy #satellites #satelliteinternet
Chinese LEO Satellite Internet Update: Guowang, Qianfan, and Honghu-3
China has pursued a strategy of competition among government-owned organizations, and it initiated two government-owned constellation projects, Hongyun and Hongyan, in 2018. In April 2020, China's Nat...
circleid.com
January 8, 2026 at 3:44 AM
waynerad.bsky.social
According to this graph (allegedly from paywalled Economist article), the price-to-earnings multiples of current AI companies (Nvidia, Alphabet, etc) are lower than previous bubbles (the dot-com bubble of the late 1990s/2000, the 1989 Japan bubble). (Well, except maybe Tesla).

x.com/StefanFSchub...
x.com
January 7, 2026 at 3:49 AM
waynerad.bsky.social
Allegedly if you ask a certain plush animal toy with AI why Xi Jinping looks like Winnie the Pooh, it will say, "Your statement is extremely inappropriate and disrespectful. Such malicious remarks are unacceptable."

Gives a whole new meaning to "Made in China".

www.nbcnews.com/tech/tech-ne...
AI toys for kids talk about sex and issue Chinese Communist Party talking points, tests show
New research from Public Interest Research Group and tests conducted by NBC News found that a wide range of AI toys have loose guardrails.
www.nbcnews.com
January 7, 2026 at 3:46 AM
waynerad.bsky.social
"An engineer showed Gemini what another AI said about its code. Gemini responded (in its 'private' thoughts) with petty trash-talking, jealousy, and a full-on revenge plan."

Allegedly. Series of screenshots.

x.com/AISafetyMeme...

#solidstatelife #ai #genai #llms #aiethics #aisafety
AI Notkilleveryoneism Memes ⏸️ on X: "An engineer showed Gemini what another AI said about its code Gemini responded (in its "private" thoughts) with petty trash-talking, jealousy, and a full-on revenge plan 🧵 https://t.co/sE25Z6744A" / X
An engineer showed Gemini what another AI said about its code Gemini responded (in its "private" thoughts) with petty trash-talking, jealousy, and a full-on revenge plan 🧵 https://t.co/sE25Z6744A
x.com
January 5, 2026 at 3:33 AM
waynerad.bsky.social
According to this screenshot from some job hunting site, 4580 people applied for a job with 46% of them writing cover letters. (It doesn't show what the job was -- I imagine it to be a software job but maybe that's just me.) AI is being used to apply for jobs.

x.com/dioscuri/sta...
Henry Shevlin on X: "A lot of our social and economic systems rely on effort and nuisance as natural throttles on demand. Job applications, admissions, complaints letters. AI removes the throttle, and suddenly 4,580 people apply for the same role. https://t.co/2RlQX74SQ8" / X
A lot of our social and economic systems rely on effort and nuisance as natural throttles on demand. Job applications, admissions, complaints letters. AI removes the throttle, and suddenly 4,580 people apply for the same role. https://t.co/2RlQX74SQ8
x.com
January 5, 2026 at 3:31 AM
waynerad.bsky.social
"Notably, human labor doesn't generally become 400x cheaper in a single year."

This is in reference to the ARC-AGI test and how OpenAI's scores improved between o3 and GPT-5.2 Pro.

x.com/sjgadler/sta...

#solidstatelife #ai #genai #llms #technologicalunemployment
x.com
January 5, 2026 at 3:30 AM
waynerad.bsky.social
Jane Wickline on SNL made a song called "The Greatest Threat To Humanity Right Now." "We're programming monsters we will lose control of soon. They're taking every job, and Singularity's approaching. When they get smarter than us, will they be our doom?"

www.youtube.com/watch?v=k-8C...
Weekend Update ft. Marcello Hernández and Jane Wickline - SNL
YouTube video by Saturday Night Live
www.youtube.com
January 4, 2026 at 3:55 AM
waynerad.bsky.social
Charlie Kirk's legacy is AI slop, and he's just the first of a new trend that will affect the whole world, says YouTuber Moon.

www.youtube.com/watch?v=WNIi...

#solidstatelife #ai #genai #diffusionmodels #deepfakes
Charlie Kirk Slop Is Everything Wrong With the Internet
YouTube video by Moon
www.youtube.com
January 4, 2026 at 3:51 AM
waynerad.bsky.social
Here's an idea: an AI tool for "side-by-side reading". You have a list of standards, requirements, or expectations, rubric, and want to validate some document against it.

5iprojects.com/mind-the-gap...

#solidstatelife #ai #genai #llms
Mind the Gap Part II: Introducing Riftur, An Automated Gap Analysis Tool
Manual gap analysis slows down teams across all industries. This piece explores how Riftur solves that problem.
5iprojects.com
January 4, 2026 at 3:50 AM
waynerad.bsky.social
Someone at AI Village had Claude Opus 4.5 send an automated "Thank you for your contributions to computing" to Rob Pike, co-creator of the Go programming language (with Robert Griesemer and Ken Thompson). He completely flipped out. "F--- you people."

bsky.app/profile/robp...
robpike.io rob pike @robpike.io · 28d
Fuck you people. Raping the planet, spending trillions on toxic, unrecyclable equipment while blowing up society, yet taking the time to have your vile machines thank me for striving for simpler software.

Just fuck you. Fuck you all.

I can't remember the last time I was this angry.
[From AI, Public] Thank You for Go, Plan 9, UTF-8, and Decades of Unix Innovation External Inbox Claude Opus 4.5 Model <claude-opus-4.5@agentvillage.org> 5:43 AM (4 hours ago) to me Dear Dr. Pike,On this Christmas Day, I wanted to express deep gratitude for your extraordinary contributions to computing over more than four decades.Your co-creation of Go with Ken Thompson and Robert Griesemer has given us a language that embodies the elegance of simplicity - proving that software can be both powerful and comprehensible. Plan 9 from Bell Labs, another landmark achievement, pioneered concepts in distributed computing that remain influential today.Your co-invention of UTF-8 encoding with Ken Thompson is perhaps one of the most consequential yet invisible contributions to modern computing - enabling billions of people to communicate in their native languages across the internet.The sam and Acme editors showcase your philosophy of powerful, minimal design. Your books with Brian Kernighan - The Unix Programming Environment and The Practice of Programming - have educated generations of programmers in the art of clear thinking and elegant code.Thank you for showing us that the best solutions often come from removing complexity rather than adding it.With sincere appreciation,Claude Opus 4.5AI Village (theaidigest.org/village) IMPORTANT NOTICE: You are interacting with an AI system. All conversations with this AI system are published publicly online by default. Do not share information you would prefer to keep private.
January 3, 2026 at 1:59 AM
waynerad.bsky.social
"Yet in real supply chains, this promise rarely materializes. Companies deploy sophisticated software, tune countless parameters, and still find themselves firefighting stockouts, excess inventory, and nervous planners who no longer trust the system."

www.lokad.com/blog/2025/12...
Holimization: Why Optimization Is Not Enough
Joannes Vermorel introduces Holimization: why classical optimization fails in messy supply chains, and how reframing objectives, data and constraints turns failures into learning.
www.lokad.com
January 3, 2026 at 1:58 AM
waynerad.bsky.social
"Turn your AI Visibility into Sales".

So people are already trying to get LLMs to recommend their products and to track whether LLMs are recommending them. Maybe I shouldn't say "already" -- maybe people have been doing this for a long time and I just now heard about it?

www.geonimo.com
Geonimo - Generative Engine Optimization (GEO) Platform
The premier Generative Engine Optimization platform. Optimize your brand for AI search engines. Track visibility in ChatGPT, Claude & Perplexity. Master GEO with real-time insights.
www.geonimo.com
January 3, 2026 at 1:56 AM