Sign in

git remote

@gitremote.bsky.social
869 followers 3.1K following 2.7K posts

Shift Left << Anti-campist social democrat. 🌹 Trans rights are human rights. 🏳️‍⚧️ Anti anthropomorphizing LLMs.

PostsRepliesMedia
git remote @gitremote.bsky.social · 27/09/2026
Ian K. Duncan's "Sex, AI, and the Apocalypse" article you had linked to earlier expresses so well why when sex is mixed with business and friendship, refusing sex means refusing career opportunities and social membership: www.iankduncan.com/personal/202...
So I will ask the question again: can a person in this world refuse sex, decline a psychological intervention, report misconduct, or exit a relationship without threatening their career and their entire social world simultaneously?
0412
git remote @gitremote.bsky.social · 25/09/2026
Their blog post example is just a demo of a general technique. "RLHF has nothing to do with LLMs" is incorrect. ChatGPT is different from the earlier GPT foundation models, because of supervised-fine tuning (SFT) and RLHF. Post-training suppressed the more unhinged responses of pure autocomplete.
Stage 1: Pretraining

Data: Large public corpora of language and code: Internet text in multiple languages, digitized books ard papers, and open-source code repositories

Process: Train a model that, given a sequence of input text, predicts a probability distribution of the next token. The model learns grammar, facts and coding from raw text.

Model: Strong text and code completion; not yet instruction‑following or preference‑aligned.


Stage 2: Supervised Fine-Tuning (SFT)

Data: Prompt–response demonstrations that specify tone, format, and constraints.

Process: Train the model to imitate the reference answers so that it follows instructions.

Model: Follows instructions reliably; matches requested tone, format, and length.


Stage 3: Reinforcement Learning from Human Feedback (RLHF)

Data: Broad prompts for writing, QA, reasoning, coding, dialogue, and tool use.

Process: Generate several diverse candidates per prompt.

Human: Human labelers rank candidates.

Data: Convert rankings to preference pairs or partial orders for training.

Reward: Reward model predicts a scalar preference score for each prompt–output pair. Once we have this model we no longer need human graders in the loop. Note: This model can be reused and can originate from a completely different base model.

Process: Optimize the policy for higher reward while limiting Kullback–Leibler (KL) divergence from the SFT reference.

Model: A model that has been trained to satisfy the reward model. It is helpful, concise, and safe; follows instructions and meets constraints.
011
git remote @gitremote.bsky.social · 18/09/2026
Yet the EA website links to what they claim is "Our list of the most pressing world problems", which has loss of control of AI systems as the first item. You click into it and it says "Recommended - highest priority We think this is among the most pressing problems in the world."
Our list of the most pressing world problems

1. Loss of control

We may develop very powerful AI systems before we’ve built adequate control mechanisms. If such systems accumulate power while misaligned with human values, the result could be catastrophic.

2. Extreme power concentration

AI could turbocharge people’s efforts to seize and hold power. Surveillance, persuasion, and control technologies could let small groups or authoritarian leaders entrench their rule — potentially with little chance of recovery.
021
git remote @gitremote.bsky.social · 15/09/2026
Alt text added for blind people who use text-to-speech software to access the content.
Trump: “We were losing badly, but then Elon went to Pennsylvania and he knows those computers, those ballot computers, and then we won in a landslide.”  Jan 19, 2025

Elon Musk: “Without me, Trump would've lost the election and the Senate and House.”  June 5, 2025

Ashley St. Clair (mother of one of Musk's children): “Elon openly bragged to me that he rigged the election for Trump. I have all the texts and emails of him stating it, too.”  May 18, 2026
141
git remote @gitremote.bsky.social · 05/09/2026
A look inside the attempted coup in Brazil shows a familiar pattern. www.theguardian.com/world/2026/s...
He describes his former boss as someone with “significant limitations in political, institutional and administrative experience” and an “almost total imperviousness to outside opinions or information that did not conform to the ‘truths’ he had already established as his own” – citing his refusal to accept facts such as the scientific evidence supporting Covid-19 vaccination.

During the election, Bolsonaro became “entirely consumed by his own efforts to remain in office”, writes Baptista Júnior. After repeatedly casting doubt on the vote, the president ordered the armed forces to investigate the electronic voting system, but refused to accept their reports when they found no irregularities.

“He may still not believe the election result even today. Whether that is a matter of character, personality or something pathological, I don’t know,” said Baptista Júnior.
000
git remote @gitremote.bsky.social · 05/09/2026
In 2002, airport security ripped out Steve Mann's devices, injuring him. Without a fully functional navigation system, he fell twice, hit his head on a pile of fire extinguishers, and passed out. He boarded the plane in a wheelchair. wearcam.org/unplugged/14...
The security guards continued to require that he turn his machine on and of and put it through the X-ray machine while also tugging on his wires and electrodes, he said. Still not satisfied, the guards took him to a private room for a strip-search in which, he said, the electrodes were torn from his skin, causing bleeding, and several pieces of equipment were strewn about the room. 

Once his system was turned off, turned on again, X-rayed and dismantled, Dr. Mann passed the security check. When he was finally allowed to go home, some pieces of equipment were not returned to him, he said, and his glasses were put in the plane’s baggage compartment although he warned that cold temperatures there could ruin hem. 

Without a fully functional system, he said, he found it difficult to navigate normally. He said he fell at least twice in he airport, once passing out after hitting his head on what he described as a pile of fire extinguishers in his way. He boarded the plane in a wheelchair.
000
git remote @gitremote.bsky.social · 03/09/2026
Alt text added for blind people who use text-to-speech software to access the content.
Remarks @remarks・8h

Vivek Ramaswamy says antisemitism is an "attack on capitalism."
091
git remote @gitremote.bsky.social · 01/09/2026
www.politico.com/news/2025/01...
The answer, our analysis found, is that Democrats would benefit significantly, with the new great state of Canada serving as a kind of second California, a massive blue state that would hold dozens of House seats and create a huge Democratic advantage in the Electoral College.
100
git remote @gitremote.bsky.social · 21/08/2026
You selectively omitted the first part of the sentence: "Emerging evidence indicates that agential AI might validate or amplify delusional or grandiose content". www.theguardian.com/technology/2...
For his paper, Dr Hamilton Morrin, a psychiatrist and researcher at King’s College in London, analyzed 20 media reports on so-called “AI psychosis”, which describes current theories as to how chatbots might induce or exacerbate delusions.

“Emerging evidence indicates that agential AI might validate or amplify delusional or grandiose content, particularly in users already vulnerable to psychosis, although it is not clear whether these interactions can result in the emergence of de novo psychosis in the absence of pre-existing vulnerability,” he wrote.
100
git remote @gitremote.bsky.social · 18/08/2026
News organizations should know by now, year 2026, that AI is a technology, not a quasi-person with responsibility. The people or organizations who deployed the technology are exactly who is responsible for AI-caused damages. www.cbc.ca/news/enterta...
The question of who exactly is responsible for damages to people, property or companies caused by false information hallucinated by AI is one that has dogged tech giants pushing the technology.
001
git remote @gitremote.bsky.social · 18/08/2026
Proper alt text added for blind people who use text-to-speech software to access the content.
Q. Okay. Well, you had to try to understand how to be an expert witness on gas-detection systems by asking it what? 

A. Again, I wanted to see what it would say.

Q. Read the prompt, sir.

A. "As an expert witness in the gas detection loss and grinding -- loss and grinding (sic) explosion, what am I basing my standard of care on as an expert witness?" 

Q. You wrote a 30-plus page written report? 

A. Yes, sir. 

Q. That's 90 to 85 percent ChatGPT? 

A. If -- if you say so. 

Q. And signed it? 

A. Yes, sir. 

Q. Sent it to your boss and says, Man, I am proud of this? 

A. Yes, sir.
1110
git remote @gitremote.bsky.social · 13/08/2026
Proper alt text added for blind people who use text-to-speech software to access the content.
Deep Unlearning

The Rise of AI and the Radicalization of a Tech Idealist

Timnit GebruAbout The Book

Timnit Gebru, one of the fiercest voices in the world speaking truth to power, chronicles her journey as a refugee and Black woman in STEM—delivering an inspiring vision for better, human-centered, equitable technology.

Timnit Gebru has been present at nearly every inflection point in tech’s recent history. She was an engineer at Apple on the cusp of the iPhone’s first release, a PhD student at Stanford’s computer vision lab as her advisor worked on the enormous datasets so crucial to AI’s development, and the most senior Black woman on the Ethical AI team at Google as DeepMind and Open AI took over the field of artificial intelligence—where she was forced out for having raised the alarm about bias in large language models, causing global shockwaves in the tech industry.Starting with her childhood during which she was forced to flee the 1998 Ethiopian-Eritrean war, through the fraught process of achieving refugee status, Gebru shows how her early love of science and math was inextricably intertwined with her belief that these were reliably objective safe spaces. Over time, through high school in Greater Boston, college at Stanford, and jobs at Apple, Microsoft, and Google, she came to realize that perhaps some of the biases and injustices she witnessed in her life outside of academics were true about technology, too.

Over the course of her life, Gebru has seen up close how the destructive ideologies and outsized egos of the tech world have come to wreak havoc on our environment, our economies, and the lives of billions of people. She has also come to understand how we could have made a different choice every step of the way, and how we still can.

Part memoir and part manifesto, Deep Unlearning is as much about ideology as it is about technology—and about how to privilege justice in both.
1214
git remote @gitremote.bsky.social · 12/08/2026
In 2021, the Chinese government spread anti-vax propaganda against Western COVID-19 vaccines, including the one produced by Pfizer. www.reuters.com/article/us-h...
BEIJING (Reuters) - Chinese state media outlets have run a series of articles criticising Western COVID-19 vaccines in the past week, including Pfizer's, while touting China-made vaccines as safer and more accessible.
100
git remote @gitremote.bsky.social · 12/08/2026
The Chinese government spread anti-mRNA-vaccine disinformation, because mRNA vaccines are a Western invention. apnews.com/article/chin...
Senior Chinese government officials have not been shy in voicing concerns about the mRNA vaccines developed by Western drug companies. They use a newer technology than the more traditional approach of the Chinese vaccines currently in use.

In December, the director of the Chinese Centers for Disease Control, Gao Fu, said he can’t rule out negative side effects from the mRNA vaccines. Noting this is the first time they are being given to healthy people, he said, “there are safety concerns.”
150
git remote @gitremote.bsky.social · 10/08/2026
This is a great illustration of: 1. what LLMs do 2. the type of people who think "LLMs are autocomplete" is a bad-faith anti-AI insult LLMs are autocomplete. You can try out OpenAI's GPT-2 (released 2019) here: huggingface.co/spaces/mkmen... (set it to GPT-2 S for a faster response)
A search field with the label "Search country" has "Aust" as the input text. "Australia" and "Austria" appear as potential autocomplete suggestions.
110
git remote @gitremote.bsky.social · 09/08/2026
That pejorative interpretation of "autocomplete" is more prevalent among those who think the term minimizes LLMs, where they think LLMs are a computational model of the human mind. Emily Bender even mentions this common misinterpretation of "stochastic parrot". spectrum.ieee.org/stochastic-p...
Another one is that “stochastic parrot” got picked up and interpreted by other people as a minimization or an insult. It was not meant that way. Other people might be using it that way, but that’s not how I intended it, because it’s just a description of what these systems actually are. To see it as an insult requires either the belief that the large language model is the kind of thing that can take offense, which it isn’t, or that these large language models should be understood as steps toward this grand ideal that I don’t hold of artificial intelligence.
1180
git remote @gitremote.bsky.social · 07/08/2026
Emily Wilson made a similar complaint. www.lrb.co.uk/the-paper/v4...
But the film wants to have it both ways. There are no gods, and yet shipwrecks and death are presented as the inevitable consequence of blinding the son of Poseidon and devouring the Cattle of the Sun. Zeus’ law is portrayed as being crucially important, and yet there is no Zeus. Nolan’s vision of the world is not quite big enough to include either real gods or real people who might be different from us.
010
git remote @gitremote.bsky.social · 28/07/2026
The video shows this graph. The regression lines for Jan 2023 to September 2024 would be roughly K-shaped for high-income versus low-income. (But this graph sucks with the uneven spacing on the x-axis.)
K-Shaped Pattern Evident for Gas Spending

Real gas cumulative growth
000
git remote @gitremote.bsky.social · 28/07/2026
For anyone who didn't read the article (not Matt), it doesn't debunk the K-shaped economy. It says the K-shaped economy for consumer spending isn't settled. (But a consumer K isn't the exact thing the other articles are talking about. If people are spending more on food, it doesn't mean it's good.)
The consumer K meanwhile remains an open question rather than a settled finding. Wealth concentration has risen on the best available real-time data, and the transmission mechanism—equity wealth effect into upper-income spending—is theoretically coherent, but it has yet to decisively show up in what distributional spending data we have. A consumer K could become more evident as wealth spillovers become more salient, or as other shocks such as geopolitical and trade tensions or slowing labor market churn has disparate effects on households. Until then, consumer bifurcation is a hypothesis worth monitoring, not a conclusion to act on.

The K, in other words, is more Wall Street than Main Street—for now.
121
git remote @gitremote.bsky.social · 28/07/2026
Here is my receipt for their presumption that Wilson is not an expert on the Odyssey. I stole this image from someone who (didn't add alt text and) posted the image to articulate their thoughts, saying, "I couldn’t have said it better myself."
Variety @Variety - 6h

"The Odyssey’ Translator Emily Wilson Slams Christoper Nolan's ‘Abysmal’ Screenplay: 'l Would Be Ashamed to Have Written Any Part of This Script’

doctor idk @bigmonkeong

all due respect but does she think she wrote the odyssey because she translated it lol

6:42 PM - Jul 27, 2026

doctor idk @bigmonkeong - 3h

obviously she can think whatever she wants but shes just so needlessly melicious in that review and full of herself but maybe thats just how all history nerds talk

translating and adapting a piece of work is something completely different so i dont think shes an authority here
010
git remote @gitremote.bsky.social · 28/07/2026
Do you mean you read Homer in various English translations, but not in various languages? Did you read her review?
Adaptations needn’t and shouldn’t follow the original with anything like the exactitude of a translation. Some of the best radically transform or resist the source material, providing what Harold Bloom called ‘strong misreadings’ of their ancestors. The most memorable creative responses to The Odyssey – such as Euripides’ Helen, Virgil’s Aeneid, Milton’s Paradise Lost, Joyce’s Ulysses, Walcott’s Omeros, Atwood’s Penelopiad, Madeline Miller’s Circe or Lisa Peterson’s play set in a modern migrant camp – reconsider the poem’s structure, narrative and values. In theory, Nolan seems to understand the assignment and its hazards.
100
git remote @gitremote.bsky.social · 27/07/2026
I'm a leftist. Elbows-up Canada is not in conflict with the aegis of the commonwealth. American identity was about separation from Britain, but for Canadians, the War of 1812 was about "us" defending Canada against US invasion. thecanadianencyclopedia.ca/en/article/t...
The ruin left behind by the British drew sharp criticism not only from the Americans, but from politicians in England and Europe. Nevertheless, the majority of Britons saw the burning of the American capital as just retaliation for the burning of York and Niagara, and it remains, however awkwardly, a point of Canadian national pride.
130
git remote @gitremote.bsky.social · 26/07/2026
"The price of food purchased for home use rose 33% from 2019 through June 2026. Consumers now are buying fewer of their favorite foods or less food overall." apnews.com/article/groc...
Grocery shopping putting a strain on family budgets

The price of food purchased for home use rose 33% from 2019 through June 2026. Consumers now are buying fewer of their favorite foods or less food overall.

Change in at home food prices, U.S. cities

Food prices have soared in recent years.

(A line graph shows food prices rose steeply from March 2020 to June 2020 and then dropped. From March 2021 to February 2023, food prices rose steeply. From February 2023 to June 2026, food prices continued to rise, but not as steeply as the earlier rises. There is a small gap in the line for October 2025.)

Seasonally adjusted; Oct. 2025 data missing due to government shutdown

Source: U.S. Bureau of Labor Statistics

Graphic: Phil Holm

AP
010
git remote @gitremote.bsky.social · 26/07/2026
Another shout-out to the news wire service Thompson Reuters, syndicated by CBC News. www.cbc.ca/news/world/m...
(lede)

Billionaire Elon Musk's hand gesture while he spoke during a celebration of U.S. President Donald Trump's inauguration drew online comparisons to a Nazi salute on Monday, but a leading tracker of antisemitism said it appeared to represent a moment of enthusiasm instead.
020
git remote @gitremote.bsky.social · 26/07/2026
AI boosters say that detractors have a "skill issue" with using AI if they experienced that generative AI's output quality is worse than doing it ourselves. The difference in experience is indeed due to a "skill issue", as we suspected. siepr.stanford.edu/publications...
In experimental settings, generative AI tools — such as chatbots and coding tools — have often been found to disproportionately improve the performance of less experienced and poorer performing workers. In one such study, researchers analyzed the impact of a generative AI assistant on customer support agents in a large call center. The assistant increased overall productivity by 15 percent, with gains highly concentrated among novice and less-skilled workers, who saw a 30 percent improvement in the number of issues resolved per hour. There was no performance improvement among highly skilled customer service agents, whose response quality fell slightly.

Other studies also show that AI tools generally speed up task completion, though the effects vary by task, context, and skill level. Figure 3 summarizes research findings on the impact of AI tools on speed across a variety of tasks. In software development, the use of GitHub Copilot — an AI tool that suggests code and functions — allowed tasks to be completed 56 percent faster, with gains concentrated among less-experienced programmers.
120
git remote @gitremote.bsky.social · 25/07/2026
It's wild that a political internet meme has convinced centrists on microblogging apps that almost nobody in the U.S. is struggling financially. The K-shaped economy has been well-documented for years by economists.
CNBC

Personal Finance

K-shaped economy is ‘alive and well,’ expert says — what new research shows

Published Fri, May 1 2026

Jessica Dickler

Key Points:

A new report by TransUnion found that as credit conditions improve for wealthier consumers, others are struggling in the face of higher costs and rising debt burdens.

Since 2023, the economy has become increasingly bifurcated, according to a new blog post by the Federal Reserve Bank of New York.

Now, consumer spending is driven mostly by high-income households, those earning more than $125,000 a year, the New York Fed researchers found.
030
git remote @gitremote.bsky.social · 25/07/2026
Statistics is on the side of leftists. The K-shaped economy means “The rich get richer, and the poor get poorer.” www.cnbc.com/2026/05/01/k...
The start of the K

The economy noticeably diverged in 2023, the New York Fed researchers found, “shortly after many of the pandemic-era subsidies for low- and middle-income households expired.”

Since then, low-income households have been hardest hit by prolonged inflation while wealth has risen fastest for those at the very top, the researchers found.
160
git remote @gitremote.bsky.social · 25/07/2026
You need to read the whole history. The K-shaped economy started after the pandemic and continued into 2026. www.cnbc.com/2026/05/01/k...
The start of the K

The economy noticeably diverged in 2023, the New York Fed researchers found, “shortly after many of the pandemic-era subsidies for low- and middle-income households expired.”

Since then, low-income households have been hardest hit by prolonged inflation while wealth has risen fastest for those at the very top, the researchers found.
000
git remote @gitremote.bsky.social · 25/07/2026
No. "Workers in retail, food service, hospitality, and lower-paid care occupations faced slower wage growth. Their sensitivity to inflation rendered them more vulnerable than workers on the higher end of the income scale." www.britannica.com/money/k-shap...
While the upper arm continued to experience growth, which segments of the economy experienced steady decline? 

Limited asset exposure. Households with minimal to no exposure to stocks, real estate, or business ownership have largely missed the asset-driven rebound. As asset prices rose, the capacity of these households to accumulate wealth remained constrained.

Low-wage service workers. Workers in retail, food service, hospitality, and lower-paid care occupations faced slower wage growth. Their sensitivity to inflation rendered them more vulnerable than workers on the higher end of the income scale.

Workers with unstable hours or gig roles. People in jobs with fluctuating schedules or in “gig economy” work tied to uneven demand were more likely to experience income volatility and heightened exposure to inflation-driven cost pressures.

Regions dependent on tourism or manufacturing. Areas that rely on tourism or cyclical manufacturing can be especially sensitive to economic downturns. When demand remains weak, recovery can be slower and uneven, prolonging employment and exacerbating income instability.

Real-world impact of a K-shaped economy

Beyond theory, a K-shaped economy has real consequences for people’s jobs, financial stability, and opportunities.

For workers

The split often shows up as a difference between job security and fragility. Depending on which jobs are in demand, workers on different arms of the K—like knowledge workers versus service and manual labor workers—experience different levels of bargaining power. Over time, these differences can affect social mobility, influencing their ability to move to higher-paid positions and overall economic status.
120
git remote @gitremote.bsky.social · 24/07/2026
Anecdote is not data. en.wikipedia.org/wiki/Food_in... Increased food insecurity from the pandemic was both real and also a worldwide phenomenon, had nothing to do with the US president.
Coronavirus (COVID-19) pandemic and food insecurity

Unemployment

Recent studies suggest that food insecurity in the United States has doubled overall and tripled among households with children since the start of the COVID-19 pandemic.[113][114] Food security rates can be predicated by the national unemployment rate because food insecurity is measured by both access to food and ability to afford it.[113] During economic downturns in the past several centuries, food insecurity and food shortages rise not only during the year of the downturn, but for several years after.[115]

In 2020, the COVID-19 pandemic and economic volatility have spurred mass layoffs or hour reductions, specifically in the transportation, service, leisure, and hospitality industries and domestic workers. As a result of lost wages, individuals and families working in these industries are increasingly more likely to be food and housing insecure.[116][117]
100
git remote @gitremote.bsky.social · 24/07/2026
Alt text added in case some people can't see the difference due to low vision.
In the original cartoon, the white man has an open mouth but no teeth showing.

In the new cartoon, the white man has an open mouth with teeth showing superimposed on to a second open mouth without teeth showing.

In the original cartoon, the black man holds the stick with both hands, his right hand over his left hand. The stick shows the letters "IND". His eye is closed and only his nose shows behind the stick.

In the new cartoon, the black man's two hands are merged into one hand. The stick shows the letters "AN". His brow is now a nose, and his nose and closed eye is now an open mouth with teeth superimposed on to a second nose.
0280
git remote @gitremote.bsky.social · 22/07/2026
No. Miriam is a good kid.
Aangad turned to Miriam, whose eyes dropped to the floor. Her voice pitched up. She waved her hands at her sides.

“That was insane,” she said. “You clearly knew everything.”

Aangad nodded, face blank, and Miriam’s family went on their way.

Aangad walked over to his parents’ table.

“What did she say?” his father asked.

“That I should have won,” Aangad said.
020
git remote @gitremote.bsky.social · 21/07/2026
ICE to Pay Thomson Reuters $125 Million to Find ‘Voter Fraud’ By 404 Media Archive Today snapshot: archive.is/2L3sU 404 Media article (subscription wall): www.404media.co/ice-to-pay-t...
404 Media

ICE to Pay Thomson Reuters $125 Million to Find ‘Voter Fraud’

Joseph Cox
Jul 17, 2026 at 11:35 AM

A new procurement record says ICE wants access to personal data, including names, Social Security numbers, and ethnicity, to investigate in part what the agency calls “voters fraud.” It also plans to use the data to investigate immigration fraud.
000
git remote @gitremote.bsky.social · 20/07/2026
If you're eating one egg a day, a 2% increase in egg costs raises your egg spending by about two dollars annually. You'd notice this? 

Please stop with the theatrics, people
5650
git remote @gitremote.bsky.social · 20/07/2026
These U.S.-focused wildfire smoke maps mislead U.S. politicians by showing Canada as a blank, uninhabited area. Canadians in Ontario breathed in smoke from the wildfires in Northeastern Ontario, but U.S.-focused news leaves it off the map. Source: www.cbsnews.com/news/wildfir... (July 19, 2026)
A color-coded map by CBS News shows areas in the U.S. northeast affected by wildfire smoke with outlines for U.S. states. U.S. areas with smoke are colored purple, U.S. areas without smoke are colored green. Canada is shown as a blank area and coded green.
100
git remote @gitremote.bsky.social · 20/07/2026
People who shop for their own groceries notice the increase in egg prices. People who have someone else buy their groceries—and only look at costs at the end the year—won't. Those who don't go grocery shopping would never know that some people remove items when they see their bill at the register.
If you're eating one egg a day, a 2% increase in egg costs raises your egg spending by about two dollars annually. You'd notice this? 

Please stop with the theatrics, people
161
git remote @gitremote.bsky.social · 19/07/2026
Alt text added for blind people who use text-to-speech software to access the content
AACR American Association for Cancer Research

Catching the Drift: How Wildfire Smoke Plumes May Increase Cancer Risk 

July 14, 2026 by Steve Schneible
010
git remote @gitremote.bsky.social · 18/07/2026
Alt text added for blind people who use text-to-speech software to access the content
cinesthetic @TheCinesthetic

Christopher Nolan's The Odyssey is projected to open with more than $200 million at the worldwide box office.

05:47 15/07/2026Big Tech Alert @BigTechAlert 

@elonmusk is no longer following @TheCinesthetic

15:51 15/07/2026
0121
git remote @gitremote.bsky.social · 17/07/2026
Elon Musk doesn't want art. He wants to endlessly critique art, because he devalues it.
Joyce Carol Oates on Twitter

So curious that such a wealthy man never posts anything that indicates that he enjoys or is even aware of what virtually everyone appreciates— scenes from nature, pet dog or cat, praise for a movie, music, a book (but doubt that he reads); pride in a friend’s or relative’s accomplishment; condolences for someone who has died; pleasure in sports, acclaim for a favorite team; references to history. In fact he seems totally uneducated , uncultured. The poorest persons on Twitter may have access to more beauty & meaning in life than the “most wealthy person in the world.”
0100
git remote @gitremote.bsky.social · 17/07/2026
Alt text added for blind people who use text-to-speech software to access the content.
Sam Stein @samstein 

This is really incredible. Docs released by Trump tonight confirm Russia tried to spread claims Biden was engaged in criminal activity vis-a-vis Burisma and that it advanced those narratives "with US officials" and planed a "high-profile corruption scandal... at the peak of the 2020 US presidential campaign.”

(Redacted) President Putin and senior Russian officials are overseeing efforts by proxies (redacted) to spread claims about former Vice President Biden as well as Ukrainian politicians and alleged Ukrainian influence in the 2016 US election. These claims include that when the former Vice President was in office, he engaged in criminal activity in his dealings with Ukraine and individuals tied to Ukrainian energy from Burisma. Derkach, Kilimnik, and other proxy actors affiliated with the Russian Government are advancing such narratives with US officials and other prominent persons as well as online, including through personal interactions and audio and documentary film releases via US and Ukrainian media outlets. 

(Redacted) These figures are conspiring to intensify their efforts as the election approaches to orchestrate a high-profile corruption scandal implicating former Vice President Biden and the Democratic Party at the peak of the 2020 US presidential campaign. Their aim is to defeat the former Vice President and ensure the President’s victory. Some of these proxy actors anticipate that Ukraine-themed narratives about Democratic corruption will play a decisive role in the election and that US persons are key to propagating these narratives.
132
git remote @gitremote.bsky.social · 16/07/2026
In computer science (a subfield of math, distinct from software engineering), sorting algorithms are at best linearithmic with respect to time, which is worse than logarithmic. Nobody educated in computer science expects a sorting algorithm to scale logarithmically.
Efficiency is a core principle of computer science. One of the first things undergraduates learn is that writing a program that sorts a list of 50 words is easy. But if you give that program 50 million words, it will likely run out of memory or take hours to finish. Much of computer science is learning the clever coding techniques that prevent this from happening. Many of these techniques take advantage of repeating patterns in the data so that as the program receives more input, it takes less time or memory to process each additional bit. Such efficiency is one reason that modern smartphones and computers are so capable and affordable. This is called logarithmic scaling, and it looks like this when you graph it:What Tech Companies Want to See When They Scale

(a graph of a log function with the Y-axis labelled average cost per user)

Large language models do not scale logarithmically. When they’re given more words to process, they get slower and use more memory—the time and resources increase faster as the input grows. In technical terms, LLMs scale quadratically. Any computer-science student knows that this is very bad.
110
git remote @gitremote.bsky.social · 16/07/2026
Software, machine learning, and math are all Science, Technology, Engineering, and Math subjects, but STEM is not one "industry" with one set of jargon. Scaling software is part of the software industry. The scaling law in ML and the log scale in math are academic observations, not industry.
The problem with generative AI, in the industry’s own jargon, is that it does not scale.
110
git remote @gitremote.bsky.social · 16/07/2026
Scaling in software means supporting a large number of users or transactions. Scaling in machine learning is the observation that deep learning models that didn't work well before start to work with more diverse training data & parameters. A logarithmic scale in math is graph of the log function.
The problem with generative AI, in the industry’s own jargon, is that it does not scale. The cost of growing from, say, a thousand users to a million is a key factor that venture capitalists examine when they evaluate start-ups. They want to see that the cost of adding each new user decreases over time, so that the company can support millions of users and make increasing profits. This is achieved partly through the careful engineering of computer systems that can efficiently handle more users who want to post photos, hail Ubers, or stream music.But the industry’s observation that bigger models tend to outperform smaller ones has given rise to a totemic belief in “scaling laws” that suggest any problem can be solved by simply making models bigger.Efficiency is a core principle of computer science. One of the first things undergraduates learn is that writing a program that sorts a list of 50 words is easy. But if you give that program 50 million words, it will likely run out of memory or take hours to finish. Much of computer science is learning the clever coding techniques that prevent this from happening. Many of these techniques take advantage of repeating patterns in the data so that as the program receives more input, it takes less time or memory to process each additional bit. Such efficiency is one reason that modern smartphones and computers are so capable and affordable. This is called logarithmic scaling, and it looks like this when you graph it:What Tech Companies Want to See When They Scale

(a graph of a log function with the Y-axis labelled average cost per user)

Large language models do not scale logarithmically. When they’re given more words to process, they get slower and use more memory—the time and resources increase faster as the input grows. In technical terms, LLMs scale quadratically. Any computer-science student knows that this is very bad.
111
git remote @gitremote.bsky.social · 15/07/2026
Alt text added for blind people who use text-to-speech software to access the content.
“Most every guy I know did what he’s being accused of at one point or another with their girlfriend at some point after a night of drinking,” complained the Donald Trump-loving comedian Adam Carolla on Sunday’s episode of his eponymous podcast. “It’s basically like: I’m drunk, I’m horny, I weigh 50 pounds more than you, we’re gonna have sex, it’s not really rape because we sleep in the same bed.”
041
git remote @gitremote.bsky.social · 12/07/2026
Proper alt text added for blind people who use text-to-speech software to access the content.
Elon Musk @elonmusk ・ 12h

He takes scamming to a whole new level x.com/cb_doge/status...

Altman @sama ・ 6h

homeboy you're the one selling public market investors on short-term space datacenters 

Elon Musk @elonmusk ・ 34m

We start flying them next year. Maybe you can come see them if your parole officer approves. 

After stealing an open source Al charity, you then stole all of Apple’s phone technology! Wow. 

What do you plan for an encore? That's tough to beat.
0102
git remote @gitremote.bsky.social · 11/07/2026
Social cryptomnesia is when the majority group takes credit from minority groups for social change. en.wikipedia.org/wiki/Social_...
Social cryptomnesia is a failure to remember the origin of a change, in which people know that a change has occurred in society, but forget how this change occurred; that is, the steps that were taken to bring this change about, and who took these steps. This may lead to reduced social credit towards the minorities who made major sacrifices that led to the change in societal values.

Minority marginalization

A minority group is defined as any group that is marginalized in the society they live in due to cultural or physical characteristics. They face unfair treatment and are discriminated against. Minority groups in a society can also be referred to as the “subordinate group”, whereas the majority of people are called the “dominant group”. The Scapegoat Theory suggests that dominant groups can position “unfocused aggression” on minorities.

Throughout history, there have been several notable social movements of minorities leading to revolutions; while the revolutions are remembered by the public, those who have been leading the change have often been forgotten. So, the societal system favors the dominant groups without giving credit to minorities. Social cryptomnesia is described as collective oblivion with social control. It causes the changes done by minority groups to be disregarded, which in turn could lead to further disparity within a society.
1120
git remote @gitremote.bsky.social · 06/07/2026
When women use AI for résumés and coding, it is seen as confirmation of incompetence. "the evaluators assumed that the AI tool had done more of the work when they believed the coder was a woman rather than a man."
The evaluators who attributed the AI-assisted resume to a woman were twice as likely to question the candidate’s competency. “She can’t even write a CV herself—not sure she has the skill to carry out the job,” said one of the evaluators of Emily’s resume.

In contrast, evaluators who attributed the AI-assisted resume to a man were twice as likely to view the candidate as showing initiative. In other words, a woman’s use of AI was evidence of inability, while a man’s use of AI to create the identical product was viewed as pragmatic problem solving.To assess why women engineers were using AI at lower rates than male engineers, the researchers asked 1,026 software engineers at the company to evaluate an identical piece of computer code. While all of the engineers reviewed the exact same code, they received varied instructions about whether the code was written with or without AI assistance, and whether the coder was female or male.

As expected, the reviewers rated the objective quality of the identical computer code similarly in all conditions. However, the reviewers offered very different ratings of the engineers who wrote the code. The reviewers imposed a competence penalty against all of the engineers who purportedly used AI as compared to non-AI users. And the penalty for AI use was much harsher for women than for men.While male AI users received 6% lower competence ratings than non-AI users, female AI users received 13% lower competence ratings, even though they all produced identical computer code. Male evaluators who did not use AI themselves imposed the most severely gender-biased competence penalties. Male non-AI users rated women engineers who used AI 26% more harshly than men who used AI to create the same work product.

The study further revealed how women’s workplace contributions get devalued when they use AI. When asked to estimate the relative contribution of the engineer versus the AI tool, the evaluators assumed that the AI tool had done more of the work when they believed the coder was a woman rather than a man. “The AI assistance is framed as a ‘proof’ of their inadequacy rather than evidence of their strategic tool use,” concluded the researchers.
000
git remote @gitremote.bsky.social · 28/06/2026
Shipping software is a social construct.
"I know it sounds extreme, but I think many engineers do not understand what shipping even is inside a large tech company. What does it mean to ship? It does not mean deploying code or even making a feature available to users. Shipping is a social construct within a company. Concretely, that means that a project is shipped when the important people at your company believe it is shipped. If you deploy your system, but your manager or VP or CEO is very unhappy with it, you did not ship."

And conversely, if you do not deploy your system to a working state but can someone make your VP or CEO happy, then you did ship. This sounds even stranger. How can you ship if the code doesn't work?
010
git remote @gitremote.bsky.social · 09/06/2026
Have you heard of "toxic masculinity" as a feminist criticism, and that it refers to the harmful ways that boys and men are expected to perform gender?
Wikipedia: Toxic masculinity

Toxic masculinity is a concept used in academic and media discussions to refer to those aspects of hegemonic masculinity that are socially destructive, such as misogyny, homophobia, and violent domination. These traits are considered "toxic" due in part to their promotion of violence, including sexual assault and domestic violence. Socialization of boys sometimes also normalizes violence, such as in the saying "boys will be boys" about bullying and aggression.

(A comic shows a man bending down to talk to a crying boy. The man's speech bubble says "The angels grew lonely, Danny, and came for your mother to take her to heaven! Now, be a brave little boy and don't cry!" The crying boy's speech bubble says "I-I won't cry, Daddy".)

(Image caption: According to social learning theory, teaching boys to suppress emotions, as in the saying "big boys don't cry", is a significant part of gender socialization in Western society.)
130
git remote @gitremote.bsky.social · 27/05/2026
I assume he blocked you because you did the exact thing he complained about.
lan Boudreau @ianboudreau.com

My favorite pro-Al argument goes like this: "In the past, other technologies have truly been transformational" and that's the whole argument
100