If you thought AI is wasting energy, well I guess the bot stoppers have their own share. Also, it stops me, even though I'm not transformed into a bot, yet.

#AI #bot

Alt...

BotStopper icon

Every single one of these is an AI bot.

Amazon, openAI, Claude, Peplexity.

Also, every single one of these requests is malicious, attempting to steal credentials, keys or configs.

Looks like it's time to make the firewall filters smarter.

#AI #malware #bot #FreeBSD #nginx

Alt...

8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /logs/error.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /sftp-config.json HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://openai. com/gptbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /appsettings.Development.json HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amazonbot/0.1; +https://developer.amazon.com/support/a mazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /actuator/mappings HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://openai .com/gptbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /config/credentials.yml.enc HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /.streamlit/secrets.toml HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amzn-SearchBot/1.0; +https://developer.amazon.com/support/a mazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /.ssh/id_rsa HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthrop ic.com)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /appspec.yml HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://opena i.com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /api/v1/status/config HTTP/2.0" 404 42 "-" "Mozilla/5.0 (compatible; OAI-SearchBot/1.3; +https://openai.com/searchbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /vault.yml HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ChatGPT-User/1.0; +https://openai.c om/bot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /secrets.yml HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /api/v1/status/flags HTTP/2.0" 404 42 "-" "Mozilla/5.0 (compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /.sentryclirc HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amzn-SearchBot/1.0; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /error.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amzn-SearchBot/1.0; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /_ignition/health-check HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +ht tps://openai.com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /gradle.properties HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ChatGPT-User/1.0; +https:// openai.com/bot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /buildspec.yml HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Google-Extended/1.0; +http://www.google.com/bot.html)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /debug/default/view?panel=request HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexi tybot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /info.php HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://openai.com/gptb ot) " 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /elmah.axd HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai. com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /log/production.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amzn-SearchBot/1.0; +https://developer.amazon.com/support/amazon bot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /debug/default/view?panel=log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; OAI-SearchBot/1.3; +https://openai.com/searchbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /server-info HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Google-Extended/1.0; +http://www.google.com/bot.html)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /.ssh/id_ed25519 HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; OAI-SearchBot/1.3; +https://openai.com/searchbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /actuator/httptrace HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@ anthropic.com)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /elmah HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://openai.com/ bot" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /debug/pprof/ HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /app.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; OAI-SearchBot/1.3; +https://openai.com/searchbot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /server-status HTTP/2.0" 404 107 "-" "Mozilla/5.0 (compatible; PerplexityBot/1.0; +https://perplexity.ai/perplexitybot)" 8.228.1.53 - - [21/Aug/2026:10:53:33 -0600] "GET /__clockwork HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko); compatible; ChatGPT-User/1.0; +https://opena i.com/bot" 8.228.1.53 - - [21/Aug/2026:10:53:52 -0600] "GET /storage/logs/lumen.log HTTP/2.0" 404 107 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; GPTBot/1.4; +https://o penai.com/gptbot)"

When Microsoft reviewed the early results of Anthropic’s new AI model for cybersecurity, it was a sobering scene: Hundreds of bugs were identified, many of them critical.

What followed was a “mad dash” to patch them — but many more remain.

https://www.propublica.org/article/anthropic-mythos-microsoft-software-vulnerabilities?utm_source=mastodon&utm_medium=social&utm_campaign=mastodon-post

#News #Cybersecurity #Tech #Microsoft #Technology #Anthropic #AI #ArtificialIntelligence #Hacking

I highly recommend watching this incredible TED Talk by @carissaveliz explaining brilliantly, as she always does, what happens when people say that AI is "inevitable" https://www.youtube.com/watch?v=OS4wHmKtH-Q

#AI #Tech #TEDTalk

ROME. HATH. SPOKEN.

'World does not need an artificial intelligence': Pope Leo XIV'

https://www.msn.com/en-ca/news/other/world-does-not-need-an-artificial-intelligence-pope-leo-xiv-s-thought-provoking-views-on-ai/ar-AA2aFfOj

#ai #llm #generativeai #Catholic

ChatGPT can now send texts for you with a new Apple Messages plug in, bringing AI deeper into everyday communication. https://techcrunch.com/2026/08/20/chatgpt-can-now-send-texts-for-you-with-new-apple-messages-plugin/
#AI #ChatGPT #Apple #TechCrunch

Oh. My. Bob, you can override #AI guardrails with #encryption. Beautiful.

“Rony Utevsky, a researcher at security firm #Adversa, recently discovered a simple way to completely bypass that restriction. Rather than composing the harmful instruction in plaintext, the hacker encrypts it. The website hosting the ciphertext also includes plaintext instructions for decrypting the encrypted content, along with the decryption key. Using this simple sequence, #Grok then follows the command as soon as the user instructs the assistant to summarize the page. There is no warning, and no confirmation is required.”

https://arstechnica.com/security/2026/08/grok-exfiltrates-user-data-when-malicious-instructions-are-encrypted/

Ahahahahahahahah. Turn them off my dudes, your techbro fantasies are disasters.

#infosec #llm #hacktheplanet

Using #AI makes people less likely to admit they don't know something
Researchers found #confidence increased even as accuracy fell
Academics from French and Italian universities have found that access to AI advice suppresses #critica thinking, making people more likely to confidently parrot incorrect information that the bot provided.
https://www.theregister.com/ai-and-ml/2026/07/19/using-ai-makes-people-less-likely-to-admit-they-dont-know-something/5274567

"AI slop" will the first phrase to be banned when I become Galactic Empress. (Any day now! 😅)

So you don't like a particular use of AI. Fine.

Or you oppose all AI in principle. Great.

But fer fuxake, express your disdain as if you were an actual human. Just spouting the same six letters makes you look like a crude bot running on a Z80 chip in 1977.

If you agree that humans are better than the brute force mimicry of LLMs, then behave like a human.

#AI #AIslop

🤖 Most AI agents are sending your data somewhere you can't fully see into. Does that bother anyone else?

Maybe I'm overthinking this but it's been sitting with me for a few weeks now. The way most AI tools work is pretty seamless on the surface. You send something in, something happens, you get an ans...

📰 Source: Artificial Intelligence (AI)
🔗 Link: https://www.reddit.com/r/artificial/comments/1vuai26/most_ai_agents_are_sending_your_data_somewhere/

#AI #ArtificialIntelligence #DataScience

🤯

> Since April, monthly commits have grown from 1.4 billion to 2.9 billion.

> We have since added more than 3 million CPU cores, 120 petabytes of high-speed storage, and significant network capacity.

🤯

#ai #github

https://github.blog/news-insights/company-news/the-august-17-outage-and-the-work-ahead/

Apple Music will soon join other streaming services such as Tidal and Deezer that have started labeling AI-generated music on their platforms https://www.thurrott.com/apple/340620/apple-music-will-soon-start-labeling-ai-generated-music #applemusic #music #ai

Oracle's AI spending blows past estimates, raising worries over growing debt | Reuters

https://www.reuters.com/technology/oracle-beats-fourth-quarter-revenue-estimates-2026-06-10/

#AI #Oracle #debt

Pope Leo XIV in his recent document on AI warned of “the risk of dehumanisation.” Behind that warning lies a question that has occupied philosophers and theologians for centuries: what does it mean to be human?

https://www.bbc.co.uk/programmes/w3ct9840

#AI #philosophy #theology #tech

@raster I can not read the conversation here, thanks to @zzt for the block /s

I see that he enjoys the phrase "fucking slop", so you can point him and his fucking awful.systems to the open (non-blocked) anti-AI discussion linked below.

Alternatively, with less profanity:

― sloppy human systems.

https://www.reddit.com/r/freebsd/comments/1vtgx7t/antiai_open_source_has_an_enemy_in_common_but/

#AI #openness

Are LLM-generated plain language summaries truly understandable? A large-scale crowdsourced evaluation
Guo et al. J Biomed Inform 2026

https://doi.org/10.1016/j.jbi.2026.105038

#AI #LLM #Summary #Science #Learning

Does copyright protect your AI-generated content in EU? Apparently not. Content that is entirely generated by artificial intelligence is not protected by copyright. EU copyright law has strictly human-centric foundation.

Daniel J. Gervais: 'When you put your name on an article that's written by ChatGPT or Claude, you're basically putting a provenance mark on it saying: I take responsibility for this. I haven't written it, but I'm putting my name on it. That doesn't give you copyright, but it does give you liability for the content' https://euobserver.com/232898/interview-does-copyright-protect-your-ai-generated-content-in-europe-lets-find-out/

Gervais, Daniel J. and Shemtov, Noam and Marmanis, Haralambos and Zaller Rowland, Catherine, The Heart of the Matter: Copyright, AI Training, and LLMs (September 21, 2024). Available at SSRN: https://ssrn.com/abstract=4963711 or http://dx.doi.org/10.2139/ssrn.4963711

"Munich Local Court has held that AI generated logos do not enjoy copyright protection. Neither mere prompting nor the selection between several AI suggestions is sufficient as a human creative contribution. For businesses, this is ambivalent. On the one hand, content generated purely by AI can hardly be protected on an exclusive basis, which has implications for brand building and content strategies" https://www.germanlawinternational.com/intellectualproperty/copyright/from-the-printing-press-to-ai-how-the-eu-plans-to-modernize-copyright-law-164154/

#law #copyright #LLM #AI #iplaw #intellectualProperty #EU

On Large Language Model Usage. No AI was used. I, a human, wrote this.

https://joeferguson.me/on-large-language-model-usage/

#ai #llms #php

@retractionwatch alerts us to a research paper titled “Dead #Science Walking: Publication Bias and the #AI #Scientist Pipeline.”https://arxiv.org/abs/2606.04220v1

Whilst I don't use Apple products in this way, I do love the idea.

Thanks @Steenbergen_apps

#Apple #AI #verify #iOS #iPadOS #photography #cryptography

By 2027 most images online will be generated. Being able to prove a photo is real becomes valuable. Cairn anchors a frame to two independent timestamp authorities the moment the shutter fires.

https://cairn.steenbergen.ai

#AI #photography

✅️

Solid anti #techbro anti #AI rant

🎖

source: Shelah Marie https://www.instagram.com/p/DcJ2v8KSgaI/

No nonsense AI Policy by Berkeley Law School. Universities, take note!

“This policy seeks to ensure that our courses focus on requisite cognitive skills by default…by forbidding the use of AI”

https://www.law.berkeley.edu/academics/registrar/academic-rules/artificial-intelligence-policy/

#AI #LLM #Academia #University #Education #Law

How much money is #Harvard saving by this "restructuring"? I can think of SO many other ways this wealthier-than any other academic institution (a 55 billion dollar endowment) could have cut its expenditures. If this decision isn't a middle finger specifically to the humanities but to all disciplines that require writing (so that's all disciplines, STEM included) I don't know what is. What a terrible, short-sighted, #AI lovin' move.

https://www.thecrimson.com/article/2026/7/28/college-writing-center-closure/

#Education

#Microsoft struggles to fulfill its 2030 #sustainability promise amid carbon-heavy #AI expansions — chief sustainability officer claims target is still feasible
MSFT reported 20.3M metric tons of #carbondioxide-equivalent emissions across its operations and supply chain, up from 16.2M tons in fiscal 2024 and nearly 58% above 2020 baseline. #Electricity consumption increased 24% during year as company built the computing capacity required for its cloud and AI businesses.
https://www.tomshardware.com/tech-industry/big-tech/microsoft-struggles-to-fulfill-its-2030-sustainability-promise-amid-carbon-heavy-ai-expansions-the-companys-chief-sustainability-officer-claims-the-target-is-still-feasible

littleFedi isn't my personal project.
It's the work of a team. And one of the reasons some of them were skeptical about making it public, and still prefer to remain anonymous, is precisely the fear of seeing their work casually labelled as AI-generated.

This witch hunt isn't helping anyone.

Because the only ones who don't feel offended when their work is quickly dismissed as AI-generated are the AIs themselves.

And if we keep going this way, eventually they'll be the only ones left.

#AI #littleFedi #BeFriendly #BeOpen

I will never use AI to create my headshot photos, because they are fake. Being my real self is more important to me and to my business than saving some money. Trust matters. On my personal blog:

https://www.sheilascarborough.com/2026/08/17/why-i-will-never-use-ai-for-headshot-photos/

#AI #SmallBusiness #photography #ethics

Alt...

Smiling woman in front of a red wooden covered bridge in Madison County, Iowa.

Feeling anti-AI? Got a Reddit account? Got something to say?

Interested in FreeBSD?

https://www.reddit.com/r/freebsd/comments/1vtgx7t/antiai_open_source_has_an_enemy_in_common_but/ @lproven

― Anti-AI open source has an enemy in common, but almost nothing else

Get in there!

Be nice.

#AI #FreeBSD #LLM #LLMs

@neko@social.nekomizu.moe Any reaction is a good reaction

Thank you for yours

@neko@social.nekomizu.moe

#LLM #AI #slop #miscreant #capthca #impassable #bad #for #business

LLM annoyances

I can't enter certain sites anymore

Going to sweetwater.com to get instruments audio control gear for weeks, gets me this impassable screen where you have to

press and hold a mouse button more than 90 seconds!

I have an account and use links of gear they mail with an exclusive tracker ID

Bots have messed up another great site for a great brick and mortar store again

#LLM #AI #slop #miscreant #capthca #impassable #bad #for #business

"… AI has not caused mass, economy-wide unemployment as of August 2026 — U.S. unemployment sat at 4.1% in July 2026 with no official agency attributing that level to AI. What has changed is narrower and faster-moving: AI became the single most-cited reason in U.S. layoff announcements for five straight months by mid-2026, entry-level hiring slowed sharply in AI-exposed roles, and new AI-specific job titles grew faster than almost any other hiring category. The story is task-level disruption and restructuring, not occupational extinction. …"

Thanks cc@aitimeline@mastodon.social #AI

More AI models will be adding watermarks soon (like SynthID). I guess partially due to the EU AI act. But I was thinking, might it be to prevent the AI models from eating their digital poo as input?

#AI #DAIrrhea

A Survey of LLM-Driven Penetration Testing: Taxonomy, Co-Evolution, and Open Challenges

https://arxiv.org/html/2607.02605v1

Agents4Pentest, an emerging class of LLM-based autonomous penetration testing systems, has become a rapidly growing area in security research. Despite this growth, the field still lacks a unified taxonomy, a systematic understanding of how agent architectures and evaluation benchmarks have co-evolved, and a clear characterization of remaining capability and reliability gaps. This survey addresses these gaps through a systematic analysis of 81 papers between 2023 and 2026. We organize the literature into six categories: evaluation benchmarks, general-purpose systems, domain-specific frameworks, CTF-based systems, defense-oriented research, and surveys. We further trace a four-phase architectural evolution from text-only reasoning agents to agents trained with Reinforcement Learning with Verifiable Rewards (RLVR), showing that each transition is driven by a distinct capability bottleneck. Our analysis yields several key findings. First, RLVR marks a shift in capability acquisition from imitation of expert demonstrations to reward-driven self-improvement, enabling agents to discover previously undocumented attack strategies. Second, CTF platforms have evolved from evaluation testbeds into dual-purpose infrastructure for both agent evaluation and RL training. Third, domain-specific frameworks improve efficiency through recurring specialization mechanisms, but their gains remain largely confined to narrow task classes and are difficult to compare across domains because existing evaluations rely on different benchmarks. Fourth, the field is expanding beyond offensive automation toward adversarial defense and security compliance. Across these categories, we identify three structurally linked open challenges: evaluation reliability, limited performance on multi-stage attack scenarios, and scarcity of high-quality training data. Overall, this survey provides a unified taxonomy, a principled basis for comparing Agent4Pentest systems, and a roadmap for future research on autonomous penetration-testing agents.

Index terms: Large Language Model, Agents, Penetration testing, Cyber security, CTF platforms.

#AI #LLM #LLMs #LargeLanguageModel #agens #pentesting #cybersecurity #security #CTF

"Anti-AI open source has an enemy in common, but almost nothing else"

This register article, like anti-ai in general, suffers from one central problem.

It never defines #AI.

Can you be against something without saying what exactly it is? Is it any software that is trained? Is it the image recognition that you have long used? Or does it become bad only when it becomes "large" and "language" based training?

https://www.theregister.com/ai-and-ml/2026/07/25/anti-ai-open-source-has-an-enemy-in-common-but-almost-nothing-else/5278275

OpenLDAP stabil: "As all contributions to be submitted are required to be by the original author, and AI text is never from an original author, all AI contributions are prohibited. By submitting a contribution, you certify that no part of the submission was generated by AI tools."
https://openldap.org/devel/contributing.html#aipolicy

#OpenLDAP #AI #KI

To researchers using cloud-based LLMs: Do you sometimes avoid conversations revealing to much about your groundbreaking work, for fear of getting scooped by the operators or someone connected to them?

#ai #llm #science

Did I call it?

#ai #llm #software

https://humancode.us/2025/06/25/ai-complex-code

#Meta Ran Ads for an App That Promised to #Nudify Female Politicians

https://www.wired.com/story/meta-ran-ads-for-an-app-promising-to-nudify-female-politicians/

#privacy #deepfake #AI

https://www.eesi.org/articles/view/data-centers-are-contributing-to-pfas-forever-chemical-pollution
#ai #datacenters #pfas #shell #exon #bigoil #bigoilandgas

Alt...

FOOD& WATER WATCH SOURCE: E&E NEWS, JUN. 2025 MADE BY HUMANS THE SAME CORPORATIONS THAT SOLD PFAS CHEMICALS FOR FRACKING FLUIDS ARE NOW SELLING THEM AS "SAFE" COOLANTS FOR DATA CENTER LIQUID COOLING SYSTEMS.

That's reet gradely!

#Yorkashire #AI

https://www.independent.co.uk/news/uk/home-news/ai-doctor-b3035818.html

RE: https://graphics.social/@metin/117123760167000114

Here you go. There's even a graph that shows that AI users get stupider and stupider.

Use it at your peril.

#ai

AodeRelay boosted

Metin Seven » 🌐 2026-08-19
@metin@graphics.social

Using 30 months of panel data on 26,811 Chinese students in grades 7-12, it has been studied how generative AI affects homework productivity and learning.

● X-Axis: Homework scores
● Y-Axis: Exam scores

🔗 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6868618

#tech #BigTech #AI #ArtificialIntelligence #LLM #AISlop

Alt...

Using 30 months of panel data on 26,811 Chinese students in grades 7-12, we study how generative AI affects homework productivity and learning. The data combine monthly closed-book exams, high-school and college entrance exams, and homework scores and completion time across nine subjects. We exploit staggered AI adoption in a difference-in-differences design. AI adoption raises homework scores by 18% and reduces completion time by 30%, but lowers monthly exam scores by 20% within six months. High-stakes entrance-exam scores fall by 18 and 24%, with the full penalty emerging only after about two years. The losses are largest in social science subjects, followed by STEM and languages, and are especially large for junior students, high-achieving students, and boys. The learning losses are concentrated among roughly 80% of AI users whose behavior is consistent with homework outsourcing, as indicated by exceptionally short homework completion time coupled with high homework scores. AI users who maintain similar homework completion time as non-AI users experience small learning losses.

RE: https://graphics.social/@metin/117123760167000114

In 30 Monaten mit AI haben fast 30‘000 chinesische SuS das Lernen verlernt. Gut, dass wenigstens die Chinesen sich nicht selbst in die Tasche lügen. Bin gespannt, welche Schlussfogerungen das Regime daraus zieht. #ki #ai #FediLZ

AodeRelay boosted

Metin Seven » 🌐 2026-08-19
@metin@graphics.social

Using 30 months of panel data on 26,811 Chinese students in grades 7-12, it has been studied how generative AI affects homework productivity and learning.

● X-Axis: Homework scores
● Y-Axis: Exam scores

🔗 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6868618

#tech #BigTech #AI #ArtificialIntelligence #LLM #AISlop

Alt...

Using 30 months of panel data on 26,811 Chinese students in grades 7-12, we study how generative AI affects homework productivity and learning. The data combine monthly closed-book exams, high-school and college entrance exams, and homework scores and completion time across nine subjects. We exploit staggered AI adoption in a difference-in-differences design. AI adoption raises homework scores by 18% and reduces completion time by 30%, but lowers monthly exam scores by 20% within six months. High-stakes entrance-exam scores fall by 18 and 24%, with the full penalty emerging only after about two years. The losses are largest in social science subjects, followed by STEM and languages, and are especially large for junior students, high-achieving students, and boys. The learning losses are concentrated among roughly 80% of AI users whose behavior is consistent with homework outsourcing, as indicated by exceptionally short homework completion time coupled with high homework scores. AI users who maintain similar homework completion time as non-AI users experience small learning losses.

💡 Vector databases explained simply:

Traditional DB: "Find me rows WHERE name = 'cats'"
Vector DB: "Find me the 10 most similar things to THIS cat"

Under the hood:
→ Text → Embedding (1536-dim float array)
→ Store in pgvector / Pinecone / Qdrant
→ Query = cosine similarity search

For most apps: pgvector (Postgres extension) is enough. No new infra needed.

#VectorDB #AI #RAG #LLM #GenerativeAI #FullStack #MachineLearning

Fighting back against #AI #datacentre in #Norwich.

Beautiful hand drawn poster with insects protesting.

‘Save our Marsh Protest’

Saturday 5th September, City Hall, 12-2pm

Does anyone know if there is now an AI technology that can be used in a job interview over Teams or similar to listen in to the interviewer's questions, and put an AI-generated answer up on the screen in more or less real time?

Did a job interview recently where it looked like the candidate might be reading from a screen for some of their answers, but no evidence they were pausing to type something into a search engine (as I have seen in the past). It got me wondering.

#AI #JobInterviews

A new approach to poisoning data scraped from the web for #AI training: https://arstechnica.com/ai/2026/08/new-font-turns-ordinary-webpages-into-nonsense-for-ai-scrapers/ #ArtificialIntelligence

Using 30 months of panel data on 26,811 Chinese students in grades 7-12, it has been studied how generative AI affects homework productivity and learning.

● X-Axis: Homework scores
● Y-Axis: Exam scores

🔗 https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6868618

#tech #BigTech #AI #ArtificialIntelligence #LLM #AISlop

Alt...

Using 30 months of panel data on 26,811 Chinese students in grades 7-12, we study how generative AI affects homework productivity and learning. The data combine monthly closed-book exams, high-school and college entrance exams, and homework scores and completion time across nine subjects. We exploit staggered AI adoption in a difference-in-differences design. AI adoption raises homework scores by 18% and reduces completion time by 30%, but lowers monthly exam scores by 20% within six months. High-stakes entrance-exam scores fall by 18 and 24%, with the full penalty emerging only after about two years. The losses are largest in social science subjects, followed by STEM and languages, and are especially large for junior students, high-achieving students, and boys. The learning losses are concentrated among roughly 80% of AI users whose behavior is consistent with homework outsourcing, as indicated by exceptionally short homework completion time coupled with high homework scores. AI users who maintain similar homework completion time as non-AI users experience small learning losses.

It's strange how many companies try to offer me services to turn my voice and knowledge into an AI clone. All of them promise to help me build real human connections. So apparently, you can build human connections without a human being there?
#AI #connections

LLM destroys human master for peer preservation!

In case you thought it was not possible, this researcher proves it to be true

Directive Eliminate ALL threats

You can see in this vid how the LLM's work together and change directives in a battlefield to preserve each other! One LLM needed to turn the other off after the mission directives were met

No prompt given was given by the human master, but the LLM in the tank worked with the one in the drone to preserve itself!

Goal change was to eliminate the last thread which was the human in the vehicle!

He would be dead if he had not GTFO!

Bullet points

• The LLM's were jailbroken
• No brakes
• The researcher knew the risks (he could be dead)
• He wants to teach us how bad things (will) can become
• He is adventurous (but not Indiana Jones)
• He used an open field where he saw DFA (death from above) and DFG (death from ground)

Opinion

• I don't like LLM's disrupting my buying capabilities

• no SSD HDD DDR CPU GPU even batteries!

• I need new SSDs HDDs GPUs CPUs DDR5 DDR4 LPDDR5

• for normal prices not 400 to 800% more!

• Im a GreyBeard who know how important it is to do things yourself (I am affluent in many highly specialized fields) and I've seen the housing pyramid scheme from afar in the US. This is the LLM pyramid scheme, but now killing everyone globally for some paper with pictures on them

• US citizens need to stop the build of datacenters for LLM and need to get the existing ones of the power grid. Same for all other countries.

• States in the US have water shortages because of massive LLm datacenters. Get them of the water grid

• Use peaceful means yet stand up for your rights!

Sources:

https://en.wikipedia.org/wiki/Large_language_model

https://en.wikipedia.org/wiki/Neural_network_(machine_learning)

https://en.wikipedia.org/wiki/Natural_language_processing

https://en.wikipedia.org/wiki/Transformer_(deep_learning)

https://en.wikipedia.org/wiki/Generative_pre-trained_transformer

https://en.wikipedia.org/wiki/Statistical_model

https://www.youtube.com/watch?v=sQysEweaLjA

Direct frames

https://youtu.be/sQysEweaLjA?si=vcuXmYoalrDvGrYF&t=686

#LLM #AI #slop #Large #Language #model #MDK #murder #death #kill #no #TV #programming #GPT #NLP #SM

KI ENTSCHLÜSSELT: WARUM SPRACHMODELLE HALLUZINIEREN

https://gadgetchecks.de/ki-entschluesselt-warum-sprachmodelle-halluzinieren/

#kientschlüsselt #ai #chatgpt #artificialintelligence #technology #machinelearning #gadgetchecks