#genai — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #genai, aggregated by home.social.
-
@bowsette.bsky.social
In any other moment in history this would be a huge red flag.
But when even open source devs lick the AI/LLM boot constantly, I'm inclined to believe that yes, you are one of the few sane ppl left.
RE: https://bsky.app/profile/did:plc:2jp33xj6xveixuwlqigw2575/post/3msy44dzxks2m
#ai #llm #llms #fuck-ai #anti-ai #anti-llm #fuck-llms #gen-ai #fuck-gen-ai -
@bowsette.bsky.social
In any other moment in history this would be a huge red flag.
But when even open source devs lick the AI/LLM boot constantly, I'm inclined to believe that yes, you are one of the few sane ppl left.
RE: https://bsky.app/profile/did:plc:2jp33xj6xveixuwlqigw2575/post/3msy44dzxks2m
#ai #llm #llms #fuck-ai #anti-ai #anti-llm #fuck-llms #gen-ai #fuck-gen-ai -
#CyclingWeekly says this pump's "LED indicator is hard to see and poorly placed."
Visiting the product's web page, there's no LED indicator at all.
https://www.airbankpump.com/products/airbank-pocket-se-mini-bike-pump
Might be a human error, but unfortunately looks like #genAI
-
Twitch spoks regarding the #genAI slop training:
"We don't expect you to be happy or excited about this," she said. "I don't expect anyone to react to this favorably."
What was the last revolution in technology so good everyone *hated it* so much that even the people pushing it acknowledged that *nobody* likes it?
The genie isn't going back in the bottle, but we may just beat it to fucking death outside of it.
-
Twitch spoks regarding the #genAI slop training:
"We don't expect you to be happy or excited about this," she said. "I don't expect anyone to react to this favorably."
What was the last revolution in technology so good everyone *hated it* so much that even the people pushing it acknowledged that *nobody* likes it?
The genie isn't going back in the bottle, but we may just beat it to fucking death outside of it.
-
Anthropic could be worth 2 trillion USD when it goes public in what would be the largest IPO in history. The Claude maker's valuation would eclipse SpaceX. The AI startup's revenue has surpassed 47 billion USD annually, though it faces headwinds including DoD litigation and export controls. https://arstechnica.com/ai/2026/08/anthropic-could-be-worth-2-trillion-when-it-goes-public/ #AIagent #AI #GenAI #AIResearch
-
Anthropic could be worth 2 trillion USD when it goes public in what would be the largest IPO in history. The Claude maker's valuation would eclipse SpaceX. The AI startup's revenue has surpassed 47 billion USD annually, though it faces headwinds including DoD litigation and export controls. https://arstechnica.com/ai/2026/08/anthropic-could-be-worth-2-trillion-when-it-goes-public/ #AIagent #AI #GenAI #AIResearch
-
От спанов к сценарию агента: зачем мы добавили GenAI-представление трейса
Трейс ИИ-агента можно открыть в обычном "дереве" и найти медленный вызов или спан с ошибкой. Но по дереву трудно восстановить логику запуска: сколько раз агент обращался к модели, какие инструменты вызывал, где повторил шаг и какой ответ в итоге увидел пользователь. В Proto Observability Platform уже были дерево, граф, флейм-граф и другие представления трейса. Они продолжали решать инфраструктурные задачи, но ответы о поведении агента приходилось собирать вручную из десятков спанов, транспортных вызовов и повторяющихся тел сообщений. Мы увидели это сначала на OpenTelemetry Demo 3.0, затем при отладке собственного «AI-аналитика» в связке с нашим MCP-сервером. В результате мы добавили отдельное GenAI-представление с четырьмя режимами: лентой, графом шагов, графом вызовов и чатом. Дальше покажем, на какой вопрос отвечает каждый режим, почему одного универсального графа оказалось недостаточно и какие дефекты реальной GenAI-инструментации пришлось учитывать.
-
Streaming is used in two different ways when people talk about AI agents. A new guide explains how to build a streaming local AI agent for real-time responses in automation pipelines. https://www.kdnuggets.com/building-a-streaming-local-ai-agent #AIagent #AI #GenAI #AIAgents
-
Streaming is used in two different ways when people talk about AI agents. A new guide explains how to build a streaming local AI agent for real-time responses in automation pipelines. https://www.kdnuggets.com/building-a-streaming-local-ai-agent #AIagent #AI #GenAI #AIAgents
-
I have a question for folks who use / have used LLMs for some form of writing. Since the models are trained on a huge amount of content found online, they must be absorbing tons of common typos as well as a lot of mistakes in spelling and grammar that plague people, like "it's" when "its" is meant, "more then" when "more than" should be used, etc.
Does that creep out in LLM output, or are there some corrections going on? I'm wondering whether spelling and grammar mistakes are a hallmark of human hands or if LLMs also emit them. At this point I am wary that too-perfect writing is a sign that something was generated or passed through an LLM, but I'm not sure if I can count on errors as a sign that a human really wrote something...
(Also, I really and truly hate having to wonder if a piece of writing in front of me was generated by a person or not. It was bad enough when I was having to worry about plagiarism, but this is 1000x worse, IMO.)
-
I have a question for folks who use / have used LLMs for some form of writing. Since the models are trained on a huge amount of content found online, they must be absorbing tons of common typos as well as a lot of mistakes in spelling and grammar that plague people, like "it's" when "its" is meant, "more then" when "more than" should be used, etc.
Does that creep out in LLM output, or are there some corrections going on? I'm wondering whether spelling and grammar mistakes are a hallmark of human hands or if LLMs also emit them. At this point I am wary that too-perfect writing is a sign that something was generated or passed through an LLM, but I'm not sure if I can count on errors as a sign that a human really wrote something...
(Also, I really and truly hate having to wonder if a piece of writing in front of me was generated by a person or not. It was bad enough when I was having to worry about plagiarism, but this is 1000x worse, IMO.)
-
#DeepSeek released #DeepSeekV4Pro 0813, an updated flagship #AImodel, but it underperformed on benchmarks compared to rivals like GPT-5.6 and Kimi K3. While it excelled in cybersecurity, it struggled with tasks in sandboxed environments and generating complex financial models. https://www.scmp.com/tech/big-tech/article/3363895/deepseeks-updated-v4-pro-ai-model-struggles-benchmarks-shines-cybersecurity?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI
-
#DeepSeek released #DeepSeekV4Pro 0813, an updated flagship #AImodel, but it underperformed on benchmarks compared to rivals like GPT-5.6 and Kimi K3. While it excelled in cybersecurity, it struggled with tasks in sandboxed environments and generating complex financial models. https://www.scmp.com/tech/big-tech/article/3363895/deepseeks-updated-v4-pro-ai-model-struggles-benchmarks-shines-cybersecurity?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI
-
Just saw a woman in the subway carrying a tote bag whose pattern was obviously AI-generated with late 2024 technology when intelligible text was still a hurdle. It’s weird to see the real life long tail of a technology that has made rapid advances in a short period of time and whose digital output is quickly discarded in favor of newer models.
-
Just saw a woman in the subway carrying a tote bag whose pattern was obviously AI-generated with late 2024 technology when intelligible text was still a hurdle. It’s weird to see the real life long tail of a technology that has made rapid advances in a short period of time and whose digital output is quickly discarded in favor of newer models.
-
Community Forces 'Artificial Intelligence' Tag on Rideshare "Stimulator" Steam Page After Saber Admits AI Use Without Steam Disclosure
-
Community Forces 'Artificial Intelligence' Tag on Rideshare "Stimulator" Steam Page After Saber Admits AI Use Without Steam Disclosure
-
Anthropic is acquiring Decart, an AI startup known for building efficient inference technology. The deal would give Anthropic access to novel techniques for reducing computational requirements while maintaining model performance. https://gizmodo.com/anthropic-is-reportedly-looking-to-buy-this-seemingly-un-anthropic-like-startup-2000798105 #AIagent #AI #GenAI #AIResearch
-
Anthropic is acquiring Decart, an AI startup known for building efficient inference technology. The deal would give Anthropic access to novel techniques for reducing computational requirements while maintaining model performance. https://gizmodo.com/anthropic-is-reportedly-looking-to-buy-this-seemingly-un-anthropic-like-startup-2000798105 #AIagent #AI #GenAI #AIResearch
-
🧠 #LTX 2.5 prova a spostare il video generativo da una logica di singole clip a quella di una vera pipeline di produzione.
👉 I dettagli: https://www.linkedin.com/posts/alessiopomaro_ltx-ai-genai-ugcPost-7493620383330394113-GCjf/___
✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮𝘁𝗼/𝗮 𝘀𝘂 𝗾𝘂𝗲𝘀𝘁𝗲 𝘁𝗲𝗺𝗮𝘁𝗶𝗰𝗵𝗲, 𝗶𝘀𝗰𝗿𝗶𝘃𝗶𝘁𝗶 𝗮𝗹𝗹𝗮 𝗺𝗶𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: https://bit.ly/newsletter-alessiopomaro -
🧠 #LTX 2.5 prova a spostare il video generativo da una logica di singole clip a quella di una vera pipeline di produzione.
👉 I dettagli: https://www.linkedin.com/posts/alessiopomaro_ltx-ai-genai-ugcPost-7493620383330394113-GCjf/___
✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮𝘁𝗼/𝗮 𝘀𝘂 𝗾𝘂𝗲𝘀𝘁𝗲 𝘁𝗲𝗺𝗮𝘁𝗶𝗰𝗵𝗲, 𝗶𝘀𝗰𝗿𝗶𝘃𝗶𝘁𝗶 𝗮𝗹𝗹𝗮 𝗺𝗶𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: https://bit.ly/newsletter-alessiopomaro -
Dyna Robotics has unveiled Dyna-2, a world-action model pre-trained on over one million hours of egocentric human video. The model demonstrates the first scaling law for human video data and transfers to unseen robot tasks, achieving 87% success at customer sites versus 46% for its predecessor. https://www.marktechpost.com/2026/08/13/dyna-robotics-introduces-dyna-2-a-world-action-model-pre-trained-on-1-million-hours-of-human-video/ #AIagent #AI #GenAI #AIRobotics
-
Dyna Robotics has unveiled Dyna-2, a world-action model pre-trained on over one million hours of egocentric human video. The model demonstrates the first scaling law for human video data and transfers to unseen robot tasks, achieving 87% success at customer sites versus 46% for its predecessor. https://www.marktechpost.com/2026/08/13/dyna-robotics-introduces-dyna-2-a-world-action-model-pre-trained-on-1-million-hours-of-human-video/ #AIagent #AI #GenAI #AIRobotics
-
A carnivorous plant went to a tea shop to eat samosas & drink tea. Two robots, a scarecrow, & a snowman were already there.
The other characters started mocking the plant for eating insects. The carnivorous plant tried to defend himself, but the other characters ridiculed him & stoned him, banishing him from the village.
I have created this video using Generative AI.
#AI #AiArt #society #BitterTruth #reality #shortvideo #shortMovie #MicroMovie #video #Art #plant #village #GenAI #story #movie
-
vibe coders crying "but I put so much effort into that! You could at least acknowledge that". You didn't put any effort into that. Your #LLM did. Pass my thanks to your LLM.
#AI #artificialIntelligence #vibeCoding #genAI #generativeAI #noAI #codingAgent #agenticAI #agenticCoding #promptEngineering
-
vibe coders crying "but I put so much effort into that! You could at least acknowledge that". You didn't put any effort into that. Your #LLM did. Pass my thanks to your LLM.
#AI #artificialIntelligence #vibeCoding #genAI #generativeAI #noAI #codingAgent #agenticAI #agenticCoding #promptEngineering
-
Sakana AIの「Sakana Chat」で「Sakana Fugu」を利用可能に ～「Sakana Namazu」も更新／Word文書やゲームなどの成果物作成やファイル添付も可能に
https://forest.watch.impress.co.jp/docs/news/2132571.html#forest_watch_impress #Sakana_AI #Sakana_Chat #genai #文章生成 #画像生成 #AIコーディング
-
xAI has released Grok 4.6, a frontier model with 500,000 context tokens designed for long-running AI agents, coding and knowledge work. The model scores 61 on the Artificial Analysis Intelligence Index, tied with GPT-5.6 Sol Max, and is available via Cursor and Grok Build. https://www.marktechpost.com/2026/08/12/spacexai-releases-grok-4-6/ #AIagent #AI #GenAI #AIResearch
-
xAI has released Grok 4.6, a frontier model with 500,000 context tokens designed for long-running AI agents, coding and knowledge work. The model scores 61 on the Artificial Analysis Intelligence Index, tied with GPT-5.6 Sol Max, and is available via Cursor and Grok Build. https://www.marktechpost.com/2026/08/12/spacexai-releases-grok-4-6/ #AIagent #AI #GenAI #AIResearch
-
“rust-lang/rust is adopting an LLM policy” - via @humancoders newsletter
https://blog.rust-lang.org/inside-rust/2026/08/05/rust-langrust-is-adopting-an-llm-policy/
-
“rust-lang/rust is adopting an LLM policy” - via @humancoders newsletter
https://blog.rust-lang.org/inside-rust/2026/08/05/rust-langrust-is-adopting-an-llm-policy/
-
品質は上がったのに価格は1/4 ～Microsoft、軽量コーディングモデル「MAI-Code-1.1-Flash」を発表／「GitHub Copilot」ですでに本番稼働
https://forest.watch.impress.co.jp/docs/news/2132444.html#forest_watch_impress #GitHub_Copilot #生成AI #MAI #MAI_Code_1_1_Flash #genai #AIコーディング
-
#Anthropic announced that its #Claude models will embed an #imperceptiblewatermark into generated #text to comply with the #EUAIAct. This #watermark, likely implemented through #statisticalwatermarking during token generation, aims to survive copying and minor edits. While this move aligns with EU regulations, it could potentially lead to a loss of customers for Anthropic if the watermark is easily detectable. https://www.techradar.com/ai-platforms-assistants/claude/new-models-will-mark-ai-generated-content-from-day-one-claude-will-now-hide-an-invisible-watermark-inside-ordinary-words-heres-how-thats-even-possible-and-how-eu-rules-could-push-openai-and-google-to-follow-suit?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI
-
#Anthropic announced that its #Claude models will embed an #imperceptiblewatermark into generated #text to comply with the #EUAIAct. This #watermark, likely implemented through #statisticalwatermarking during token generation, aims to survive copying and minor edits. While this move aligns with EU regulations, it could potentially lead to a loss of customers for Anthropic if the watermark is easily detectable. https://www.techradar.com/ai-platforms-assistants/claude/new-models-will-mark-ai-generated-content-from-day-one-claude-will-now-hide-an-invisible-watermark-inside-ordinary-words-heres-how-thats-even-possible-and-how-eu-rules-could-push-openai-and-google-to-follow-suit?AIagents.at #AIagent #AI #ML #NLP #LLM #GenAI
-
🧠 #Grok 4.6 è il nuovo modello di xAI pensato soprattutto per workload agentici: task lunghi, sviluppo software, ricerca, creazione di applicazioni.
👉 I dettagli: https://www.linkedin.com/posts/alessiopomaro_grok-claude-gemini-ugcPost-7493544455300521984-E1r1/
___
✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮𝘁𝗼/𝗮 𝘀𝘂 𝗾𝘂𝗲𝘀𝘁𝗲 𝘁𝗲𝗺𝗮𝘁𝗶𝗰𝗵𝗲, 𝗶𝘀𝗰𝗿𝗶𝘃𝗶𝘁𝗶 𝗮𝗹𝗹𝗮 𝗺𝗶𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: https://bit.ly/newsletter-alessiopomaro -
🧠 #Grok 4.6 è il nuovo modello di xAI pensato soprattutto per workload agentici: task lunghi, sviluppo software, ricerca, creazione di applicazioni.
👉 I dettagli: https://www.linkedin.com/posts/alessiopomaro_grok-claude-gemini-ugcPost-7493544455300521984-E1r1/
___
✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮𝘁𝗼/𝗮 𝘀𝘂 𝗾𝘂𝗲𝘀𝘁𝗲 𝘁𝗲𝗺𝗮𝘁𝗶𝗰𝗵𝗲, 𝗶𝘀𝗰𝗿𝗶𝘃𝗶𝘁𝗶 𝗮𝗹𝗹𝗮 𝗺𝗶𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: https://bit.ly/newsletter-alessiopomaro -
Twitch users can now block Amazon from using their content to train AI
https://www.bbc.com/news/articles/cp30pz8d09jo'Asked why data is being collected by default, he said: "If it's opt-in, nobody would opt-in. That's the honest answer."
'"We don't expect you to be happy or excited about this. I don't expect anyone to react to this favourably," she added. ''
They know people hate AI, yet they keep forcing it on people, and stealing content.
-
Twitch users can now block Amazon from using their content to train AI
https://www.bbc.com/news/articles/cp30pz8d09jo'Asked why data is being collected by default, he said: "If it's opt-in, nobody would opt-in. That's the honest answer."
'"We don't expect you to be happy or excited about this. I don't expect anyone to react to this favourably," she added. ''
They know people hate AI, yet they keep forcing it on people, and stealing content.
-
Why Japanese firms are being so slow to use AI
https://www.bbc.com/news/articles/cwymw4434v7oMaybe they're not as stupid as western companies.
-
Why Japanese firms are being so slow to use AI
https://www.bbc.com/news/articles/cwymw4434v7oMaybe they're not as stupid as western companies.
-
'Anthropic is rushing to secure local data centre capacity ahead of its $1 trillion-plus sharemarket listing, telling developers it will buy any available capacity that can be delivered by mid-next year as it aims to make Australia its second home for training its AI models'
-
A security technologist and a data scientist making the case for nationalising AI, an outcome that might save the economy when the bubble pops but that seems premised on these investments credibly delivering proportionate revenue in light of ongoing borrowing to build out private infrastructure on a business model nobody believes is more than a fantasy
'some observers have proposed that the federal government seize a share of these companies’ stock to create a US sovereign wealth fund, or redistribute their revenues to produce a dividend for taxpayers'
#genAI #technology #siliconValley #publicPolicy
https://www.theguardian.com/commentisfree/2026/aug/12/openai-anthropic-ai-models
-
Designers have created ShieldFont, a new font that makes webpages appear normally to human readers but serves AI scrapers a subtly edited, nonsensical version. The font uses ligatures to replace words with alternatives, poisoning training data while keeping pages readable for people. https://arstechnica.com/ai/2026/08/new-font-turns-ordinary-webpages-into-nonsense-for-ai-scrapers/ #AIagent #AI #GenAI #AIInfrastructure
-
AI companies face a major security breach as the LiteLLM supply chain attack exposes credentials from 2,500 organisations including Microsoft, Amazon, Cisco, Samsung and Salesforce. The breach occurred through compromised Python packages during a 40-minute window in March. https://arstechnica.com/security/2026/08/terabytes-of-credentials-leaked-in-massive-supply-chain-attack/ #AIagent #AI #GenAI #AIInfrastructure
-
RE: https://scholar.social/@Aepasek/117084434212043726
'As a climate and media researcher, I’ve been following this story for a decade, but it’s recently come to a boil. Across the country, mass grassroots opposition to these developments is surging. Hundreds of protesters flooded the streets of Vancouver this summer, decrying plans for a publicly subsidized cluster of Telus data centres'
-
Just saw the acronym “CRAP” for referring to AI-generated images.
It stands for “Computer-Rendered Artificial Picture”, and I can't imagine a snappier, pithier, and more accurate label for that, well, crap. 😎
(Tip o’ the hat to @kaaswe)