It happened.

“Can you help me debug this program? I vibe coded it and I have no idea what’s going on. I had no choice – learning this new language and frameworks would have taken ages, and I have severe time constraints.”

Did I say “no”? Of course not, I’m a “nice guy”. So I’m at fault as well, because I endorsed this whole thing. The other guy is also guilty, because he didn’t communicate clearly to his boss what can be done and how much time it takes. And the boss and his bosses are guilty a lot, because they’re all pushing for “AI”.

The end result is garbage software.

This particular project is still relatively small, so it might be okay at the moment. But normalizing this will yield nothing but garbage. And actually, especially if this small project works out fine, this contributes to the shittiness because management will interpret this as “hey, AI works”, so they will keep asking for it in future projects.

How utterly frustrating. This is not what I want to do every day from now on.

AI Browser Dia Launches Publicly on Mac
The Browser Company’s Dia app is now open to anyone on Mac. It’s the first time the AI-powered browser has been widely available since its beta launch in June.

Image

Following on from Opera’s Neon, which arrived last month, Dia is another AI-first browsing experience that’s centere … ⌘ Read more

Arduino UNO Q Combines Qualcomm Dragonwing QRB2210 and STM32 MCU
Qualcomm Technologies has announced its plan to acquire Arduino, marking a major development in the embedded and maker ecosystems. The acquisition aims to combine Qualcomm’s edge computing and AI expertise with Arduino’s large developer community and open hardware approach. As part of the announcement, both companies introduced the Arduino UNO Q, a new dual-processor development […] ⌘ Read more

Live: ASX to open steady as Wall Street slips from record highs
Gold prices surge to a new record, above $US4,000 an ounce, as the US government shutdown drags on past its first week, and the Australian share market is likely to open flat after Wall Street’s AI tech rally takes a breather. ⌘ Read more

Can private information uploaded to ChatGPT be found by other users?
Data experts say it is hard to know what the implications are for 3,000 flood victims who have had personal information uploaded to the AI platform by a government contractor. ⌘ Read more

How GitHub Copilot enabled accessibility governance process improvements in record time
See how we turned weekly accessibility grade signals into an automated, accountable remediation workflow—powered by GitHub Copilot and cross‑functional collaboration.

The post [How GitHub Copilot enabled accessibility governance process improvements in record time](https://github.blog/ai-and-ml/github-copilot/how-we-automated-accessibility-compliance-in-five-h … ⌘ Read more

Deloitte erstattet Australien Geld zurück
Das Beratungsunternehmen Deloitte muss der australischen Regierung für einen fehlerhaften Bericht Geld zurückerstatten. Das vom Arbeitsministerium für umgerechnet knapp 250.000 Euro in Auftrag gegebene Dokument wurde offenbar zumindest teilweise per KI generiert und enthielt falsche Fußnoten und Zitate. Der peinliche Vorfall verweist erneut auf die Tücken KI-generierter Texte und Bilder, die zunehmend als „KI-Gatsch“ („AI Slop“) bezeichnet werden. ⌘ Read more

Robin Williams’s daughter begs fans to stop making AI content of her father
The daughter of the late actor Robin Williams has condemned AI-generated content featuring her father’s likeness. ⌘ Read more

The developer role is evolving. Here’s how to stay ahead.
AI is changing how software gets built. Explore the skills you need to keep up and stand out.

The post The developer role is evolving. Here’s how to stay ahead. appeared first on The GitHub Blog. ⌘ Read more

Radxa Announces Fogwise AIRbox Q900 for Industrial Edge AI
Radxa has announced the Fogwise AIRbox Q900, a rugged edge AI system powered by Qualcomm’s IQ-9075 processor. The compact unit delivers high-performance compute with industrial reliability, targeting real-time inference in manufacturing, robotics, smart cities, and research. The AIRbox Q900 is powered by the Qualcomm IQ-9075 SoC. It integrates an octa-core Kryo Gen 6 CPU based […] ⌘ Read more

Apple’s 2026 Smart Glasses: Five Key Features to Expect
Apple is working on a set of smart glasses to rival the Meta Ray-Bans, and now that Meta has debuted glasses that include a display, Apple wants to speed up development on its first-generation model. Work has stopped on the next Vision Pro so that Apple can prioritize getting the glasses to market.

![](https://images.macrumors.com/article-new/2021/01/Apple-Glasses-Yellow-Featur … ⌘ Read more

[$] LWN.net Weekly Edition for October 2, 2025
Inside this week’s LWN.net Weekly Edition:

• Front: Fedora and AI; Linting kernel Rust; openSUSE Leap 16; mmap() file operation; 6.17 statistics; dirlock.

• Briefs: Bcachefs removal; Alpine /usr merge; F-Droid; Fedora AI policy; OpenSUSE Leap 16; PostgreSQL 18; Radicle 1.5.0; Quotes; …

• Announcements: Newsletters, conferences, security updates, patches, and more. ⌘ Read more

[$] Fedora floats AI-assisted contributions policy
The Fedora \
Council began a process to create a policy on AI-assisted
contributions in 2024, starting with a survey to ask the community
its opinions about AI and using AI technologies in Fedora. On
September 25, Jason Brooks published
a draft policy for discussion; so far, in keeping with the spirit of
compromise, it has something … ⌘ Read more

Spec-driven development: Using Markdown as a programming language when building with AI
I coded my latest app entirely in Markdown and let GitHub Copilot compile it into Go. This resulted in cleaner specs, faster iteration, and no more context loss. ✨

The post [Spec-driven development: Using Markdown as a programming language when building with AI](https://github.blog/ai-and-ml/generative-ai/spec-driven-development-using-markdown-as-a-p … ⌘ Read more

AI & LLM Addiction In School ⌘ Read more

AI Therapy & The Risks of Cheap Mental Health ⌘ Read more

AI’s Impact on Intimate Relationships ⌘ Read more

MSI EdgeXpert Compact AI Supercomputer Based on NVIDIA DGX Spark
The MSI EdgeXpert is a compact AI supercomputer based on the NVIDIA DGX Spark platform and Grace Blackwell architecture. It combines a 20-core Arm CPU with NVIDIA’s Blackwell GPU to deliver high compute density in a 1.19-liter form factor, targeting developers, researchers, and enterprises running local AI workloads, prototyping, and inference. The EdgeXpert achieves up […] ⌘ Read more

Using AI to map hope for refugees with UNHCR, the UN Refugee Agency
With the help of GitHub, UNHCR turned drone imagery into maps — helping refugees in Kakuma and Kalobeyei build sustainable, powered communities.

The post Using AI to map hope for refugees with UNHCR, the UN Refugee Agency appeared first on [The GitHub Blog](https://github. … ⌘ Read more

A step-by-step guide to modernizing Java projects with GitHub Copilot agent mode
Learn how to use GitHub Copilot agent mode to modernize legacy Java projects with guided upgrades, automated fixes, and cloud-ready migrations.

The post A step-by-step guide to modernizing Java projects with GitHub Copilot agent mode … ⌘ Read more

Gartner positions GitHub as a Leader in the 2025 Magic Quadrant for AI Code Assistants for the second year in a row
Our commitment is to empower every developer and stay true to our north star by building an open, secure, and AI-powered platform that defines the future of software development.

The post [Gartner positions GitHub as a Leader in the 2025 Magic Quadrant for AI Code Assistants for the second yea … ⌘ Read more

Se tens conta LinkedIn e não queres que usem os teus dados e posts para alimentar as suas IAs, dá para desativar aqui: https://www.linkedin.com/mypreferences/m/settings/data-for-ai-improvement

Since Google announced their intentions to heavily limit sideloading on Android, starting end of 2026, I’ve been looking for potential solutions, for this policy change, that threatens the majority of projects I maintain, in some way. Google already killed my browser project years ago, but I have no other choice, than to fight this, any way I can.

The best choice to deal with this, will probably be the Android Debug Bridge, which can be used not only to install apps unrestricted, but also to uninstall, or remove, almost any unnecessary part of the OS. Shizuku, combined with Canta Debloater, is the winning combination for now.

I’ve already removed most Google apps from my device: the annoying AI assistant, the stupid Google app adding the annoying articles, left of your homes screen, Google One, Gboard, Safety app… it’s amazing, no distracting Google slopware, like in the good old Android 2 days! And I absolutely intend to keep it this way, from now on, no new Google apps or services on my devices, unless Google can give me a good enough reason, to allow them there and whenever the app that verifies signatures, to block installing apps not approved by Google, I’ll just remove it from my device and advocate others do so too.

Observador a fazer debates autárquicos, a convidar CH mas não BE ou Livre (vem, convidaram-nos em Almada, mas só aí… devia ser porque estavam a contar com eles para fazer oposição ao PS)… acho que sim, faz todo o sentido, não vale a pena esconder ao que vão.

#observador #ptpol #autarquicas

How about no longer using in-browser Git repo viewers? Make the AI bots do the work and actually clone the repo.

@movq@www.uninformativ.de Interesting, yes. I didn’t know that.

No AI being used is really great. However, the same clips shown over and over again and some images being mirrored was quite annoying to me. Also, there were some quite terrible computer animations and sometimes the narration and picture didn’t match at all. Talking about the medieval period and then showing an image from the 18th hundred or so. What the heck?

These production issues made me sceptical pretty much early on. So I quickly crosschecked Wikipedia. But it seems spot on from what I’ve read. Very good. Also, the narrator’s voice was really nice to listen to.

Eels are fascinating creatures. :-)

Hmm, gnu.org is slow as heck. Shorter HTML pages load in about ten seconds. This complete AWK manual all in one large HTML page took a full minute: https://www.gnu.org/software/gawk/manual/gawk.html Is there maybe some anti AI shenanigans going on?

In any case, I find the user guide super interesting. My AWK skills are basically non-existent, so I finally decided to change that. This document is incredibly well written and makes it really fun to keep reading and learning. I’m very impressed. So far, I made it to section 1.6, happy to continue.

Acreditar só nas coisas que corroboram as nossas opiniões é um tipo de viés difícil de escapar…
dito isso… rsrsrsrsr

“Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task”
https://arxiv.org/abs/2506.08872

Update: Puz o link direto pro artigo, que o divulgador inical parece suspeito (vide replies)

Segunda notícia que leio do mesmo jormal, mesmo dia, sobre a #AmálIA.

https://eco.sapo.pt/2025/09/02/equipa-de-peritos-esta-a-analisar-possiveis-impactos-legais-do-amalia-incluindo-nos-direitos-de-autor/

Este pôs-me ligo a resmungar com o título - e não melhorou.

O título é: “Equipa de peritos” está a analisar “possíveis impactos legais” do Amália, incluindo nos direitos de autor

Pôs-me a resmungar porque o projecto até já era para ter sido lançado, agora tem data de lançamentk para setembro (sim, este mês) e afinal… ainda nem fizeram aquele que devia ser o primeiro passo? Então e se agora afinal os dados não podem ser usados? Ou uma parte deles - vão tirá-los da base de dados iniciais e retreinar os modelos?

A questão é tão óbvia que até os jornalistas se lembraram de a fazer. E aí é que comecei mesmo a resmungar. O responsável pelo projecto podia ter dito “a legislação não impossibilita o início dos trabalhos.” Porque claro que não impossibilita. Mas o problema é que iniciar os trabalhos com todos os dados sem saber quais é que vão ser excluídos pode até ser contraproducente. E se afinal nenhum pode? Lá se foram os 5.5M€ que a brincadeira custou?

Mas a resposta foi pior, foi “Sendo um projeto público, desenvolvido em ambiente de investigação e seguindo um modelo de código aberto, a legislação não impossibilita o início dos trabalhos.”

½

Image

@prologic@twtxt.net I’m doing that now as well, but I don’t think this is a good solution. This is going to hurt “self-hosting” in the long run: I cannot afford true self-hosting where I actually do host everything here at home – instead, I must use a cloud provider / VPS for that. It is only a matter of time until my provider starts doing AI shit as well (or rather, the customers do it) and then what? I get blocked, e.g. I can’t send email to (some) people anymore. This is already bad and it’s going to get worse.

Hahaha, how funny is that!? The Dunning-Kruger effect research was sparked off by two bank robbers who rubbed lemon juice in their faces as this makes them invisible, just like invisible ink. :‘-D https://en.wikipedia.org/wiki/1995_Greater_Pittsburgh_bank_robberies

I now imagine all these AI wankers with lemon juice in their faces.

“But all your stuff is MIT licensed! They are allowed to do that!”

Haha. As if they would care. They crawl everything they get their hands on.

Besides, that’s not true, the license states that the copyright notice must be retained. “AI” breaks that. They incorporate my code and my articles in their product and make it appear as if it was their work.

Why do I care about this?

1. The load will become a problem at some point.
   
2. These crawlers and the current “AI” in general are breaking the rules. I am supposed to be paying for every little thing, I get sued for “piracy”. But apparently, these rules only apply to me. If I had more money, I could break them. Fuck that.
   
3. I simply don’t want it. Period.

@thecanine@twtxt.net We don’t use Microsoft at work – but similar products of other big companies. They’re all doing the same. The core product gets worse and worse, because they focus so much on vomiting “AI” over everything.

It will die down eventually. I hope.

We use all the Microsoft programs at work - Teams and Outlook especially.

After all kinds of technical problems with Teams, that sometimes go unresolved for over a year, Microsoft shifted their priorities away from fixing things and towards adding an annoying AI Copilot button, that just takes up space and all it does, is loads the website in Teams, so I disabled it. Soon they just add it back, but in a different row of icons, therefore it’s now a different button, you have to disable (I think they added yet another one, to the Teams, on my work phone and I had to disabled that too). Not too long after, the desktop one just enabled itself, because of “an error” and I can disable it, but doing so activates a popup, that begs you to turn it back on, every once in a while. You can’t disable the popup and can only click “Yes” or “Not now” on it. I still keep it disabled, out of principle, but yesterday I noticed yet another Copilot button, this time in the top right corner of my Outlook and this one cannot be disabled, on the business version of Outlook and even on the personal one, it’s only possible to do it through hidden privacy settings, by prohibiting the program from connecting to Microsoft servers, for extra “features”.

There’s people complaining about it online, so it’s clear nobody really wants it, but at this point Microsofts position is that you will have at least one useless AI button on your screen, at any given time, and you will be happy. And yes, their AI sucks and if I absolutely have to use AI for something, there’s already 2 better options, we have access to, at work.

James Gleick: “The lie of AI”
https://around.com/the-lie-of-ai/
Long read, it starts with Claude Shannon and Markov chains…

A very good one:
“Whatever AI Looks Like, It’s Not”

(notice the photo by Thomas Schmitt that illustrates the article)

https://archive.ph/I4nRA

Another wave of tens of thousands of hints by the same bot on the same file:

https://movq.de/v/61f8d39d2f/s.png

There’s probably a simple explanation for this: Maybe this bot was written with “AI” and it’s simply complete garbage.

This isn’t a serious threat for my low-profile website – yet. Can’t wait for this to get worse …

I intentionally use #AI tools to work…

#GitHub is becoming unusable! Images on discussions no longer load for me most of the time, posts keep loading infinitely and it breaks the reply feature… I’m really pissed by this. Is it just me? Does it think I’m a bloody AI scraper bot or something? What the hell.

Bloody AI clowns:

https://movq.de/v/34ccfc125c/s.png

How many IPs am I supposed to block, eh?

This: https://www.thebignewsletter.com/p/why-are-we-pretending-ai-is-going

Nuke it from orbit: https://www.aaron.ai/

Nooooooooooooooooooooooooooooooooooooooooooooooooooooooo, the doctors have started using AI voice agents and they understand jack shit. 😭😭😭

Gente amiga, que provider europeu recomendam para registo de domínios?

O Gandi subiu imenso os preços, e estou a olhar para o OVH, mas antes de mergulhar queria saber o que se usa por aí

AI this, AI that.

Tech is no longer interesting. I need to find a new field.

@prologic@twtxt.net

I’ve been playing around with AI at home over the past few months and building my own neural networks from scratch (in Go) with genetic algorithms

Oh, is that all 🤣

That sounds like some intensive ‘playing around’ haha

@eldersnake@we.loveprivacy.club Yeah for sure! The thing that annoys me about a lot of this, is the sheer fact you can’t really self-host let alone self-train these things I’ve been playing around with AI at home over the past few months and building my own neural networks from scratch (in Go) with genetic algorithms on a few tasks and training sets, but man it’s hard™ 🤣 I feel like we’re doing something wrong here…

How you can tell a “review post” on some random website was written by AI?

Ergonomically nicer than its binocular counterpart

How exactly is this a reason to avoid?! 🤦‍♂️

** Om nom nom LLMs, in which I respond to Simon Willison’s analogy **
I am hesitant to wade into the tumultuous waters that are the discourse around generative AI and LLMs, but this morning I came across a thing that so thoroughly melted my brain I feel uncontrollably compelled to respond.

This morning, at evidently 4:10 AM (no mention of timezone), Simon Willison shared the following blog post, quoted here in full:

Quitting programming as … ⌘ Read more

LOL #genAI

Image

“Eles que vão mas é trabalhar!”

É fácil ouvir críticas destas, ou ignorar o trabalho (voluntário, a tirar o tempo pessoal) que o associativismo dá. Há causas que defendemos, posições que temos, que são sustidas e representadas por movimentos associativos - que precisam de voluntários, associados, doações, ou de ouvir um “obrigado” de vez em quando.

Há por aí uma associação com que te revês ou que gostas? Que tal ires até eles e dares-lhe uma ajudinha? 😉

#associativismo

#Meta to the #EU: “the focus should be on creating a regulatory infrastructure that ensures any licence that is sufficiently permissive for the user is considered open source, rather than anointing specific licences as “open source”.”

Brilliant, sure, let’s ignore existing definitions and go with gut feeling (incidently, Meta has a gut feeling generator).

Source: https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/14625-Apply-AI-Strategy-strengthening-the-AI-continent/F3563576_en

#OpenSource

@movq@www.uninformativ.de I think it’s here on MIT’s website: Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task 🤔

@aelaraji@aelaraji.com i’m so sick of AI summaries they piss me tf off

FFS! Can’t I just get results, accurate no BS results? No erroneous/misleading AI-Slop of a summary I’ve never asked for ? I get it, there is plenty of people who LOooove (if not worship) that shit, Good for them! But at least make it opt-in or add in some kind of “Do Not Slop” browser option (as if the “Do Not Track” one made a difference, but I digress). Shit’s only going down-hill from here, I might as well as just spin up my own Searx instance and call it a day.

Soooo very very close! 😅

A sua “voz” alimenta uma horda inflamada de pessoas de mal com a vida, que se digladiam na sedenta procura de “culpados”: hoje os “esquerdalhos”, amanhã os imigrantes, no dia seguinte os muçulmanos, e por aí adiante. Para esta gente, a Verdade é um pequeno pormenor sem importância. Mithá Ribeiro não é um homem qualquer. É docente, com obra editada, e por isso a sua defesa reiterada do salazarismo é alarmante. Ao procurar reabilitar a memória de um regime autoritário, racista e repressivo, Mithá Ribeiro banaliza a violência do Estado Novo, branqueando décadas de censura, perseguição política, tortura, negação de liberdades fundamentais, e um País atrasado, profundamente machista, com mulheres silenciadas por uma pobreza de espírito máscula que ainda hoje se julga superior e faz vítimas.

2/n

Lately (since there are AI summaries at the top), each time I Google for the answer to a question, the AI summary has at least a part of the answer wrong. It makes up laws that do not exist, books that were never published - in sum, well written sentences that make linguistic sense, but with made up content.

Let me repeat: each time. Maybe I only search for hard stuff, or fringe stuff, or this some other explanation - but seriously, it’s hard to understand how isn’t Google ashamed of its AI overviews… or not sued under some regulation regarding fake news.

PS: yes, I know, my fault for using Google as a search engine.

Of Pointlessware and CEOs
Had a moment, to check up on some of the companies, I stopped following, get to The Browser Company and see their newest product - it’s just Chrome, with an AI chat window pop-up and that’s it. Something Canary Chrome, come with already.
I see Theo from T3.gg, making fun of it on YouTube and promoting “his” product - an AI chat app, where you can choose from multiple models, by all the popular AI companies. Something I already have a worse version of, at work and I don’t even use it.
There’s also an interview, about the future of virtual keyboards, surely this is at least actually a real thing and not more pointless horse shit. I check the website of the keyboard SDK, and it’s around 20 identical apps, that just copy the same keyboard SDK/api and slap chatgpt features on top - in the App Store, these are surrounded by chatgpt clones, that just feed the users prompts, into the real thing and put ads, next to the answers.

Depois do anúncio de um Ministério novo que tem como um dos objectivos “desburocratizar o estado”, temos as notícias sobre o derrape do PRR (alguém surpreendido?) e consequente narrativa de que é preciso desburocratizar. Só que olhando para o relatório da comissão de acompanhamento chega-se à conclusão que sim, realmente a máquina burocrática do estado não está a bem oleada, só que não é por excesso de burocracia (que está lá porque é necessária) mas sim porque a máquina do estado não tem os recursos necessários - estruturais, técnicos e humanos.

E já sabemos o que quer dizer “menos burocracia” para a AD - olhemos para a Lei dos Solos como um exemplo de como com o disfarce da simplificação burocrática o que obtivemos foi menos protecção ambiental. Vem aí mais disto - estejamos atentos.

#ptpol

GitHub for Beginners: Code review and refactoring with GitHub Copilot
Learn how to use GitHub Copilot to help review and polish your code.

The post GitHub for Beginners: Code review and refactoring with GitHub Copilot appeared first on The GitHub Blog. ⌘ Read more

Meta: Milliarden-Investition in Scale AI in Planung
Laut internen Quellen soll sich Meta in Gesprächen mit Scale AI befinden, das Datensätze für KI-Training bereitstellt. ( Meta, KI)

Image

Urban planning efficiency rises with AI, but challenges remain
New peer-reviewed research has found artificial intelligence could help boost productivity in urban planning but some experts warn government regulation is needed in the adoption of AI. ⌘ Read more

從零開始編寫 Mcp Server， 發佈上線，超簡單全網最細
MCP 簡單來說就是 AI 大模型的標準化工具箱。在我們之前的很多期視頻裏面，只是討論瞭如何使用別人已經編寫好的 MCP Server，對接到 AI 客戶端裏面，實現各種智能體與工作流。本期視頻我們轉換到開發者的角度，超簡單，幾分鐘之內編寫一個 MCP Server，並且把它發佈上線，提供給所有人去使用。本期視頻內容較多，並且十分硬核，可以幫助我們從開發者的角度更加深入理解 MCP 協議。在爬爬蝦 ⌘ Read more

Could Apple Ditch Siri Name in Major AI Rebrand at WWDC?
Apple will highlight its AI strategy at Monday’s WWDC 2025 keynote, with its much-talked-about “Liquid Glass” software redesign playing a secondary role in announcements, claims industry analyst Ming-Chi Kuo.

Image

Reports leading up to WWDC have indicated that iOS 26 will feature a [major design overhaul](https://www.macrumors.com/2025/06/06/ios- … ⌘ Read more

重磅！Qwen3 Embedding 王炸登場，AI 文本理解邁入新紀元！
各位 AI 愛好者、開發者們，準備好了嗎？今天，一個能徹底顛覆你對 AI 文本理解認知的 “王炸” 產品——Qwen3 Embedding 系列模型——正式發佈了！它不僅僅是一個模型，更是一個將帶領我們進入 AI 文本處理新紀元的里程碑！🤯💥 爲什麼說它是 “王炸”？五大亮點讓你不得不服！Qwen3 Embedding 系列模型，作爲 Qwen 家族的最新力作，專爲文本嵌入和排序任務而生，它繼承了 ⌘ Read more

On my blog: Generative AI Wish List https://john.colagioia.net/blog/2025/06/08/ai-wish-list.html #artificialintelligence #harm #rant

AI-Gigafactory: Kanzler Merz trifft Nvidia-Chef in Berlin
Jensen Huang kommt kommende Woche nach Berlin und trifft Bundeskanzler Friedrich Merz. Union und SPD wollen eine AI-Gigafactory in Deutschland fördern, die 100.000 GPUs hat. ( KI, Nvidia)

Image

Assigning and completing issues with coding agent in GitHub Copilot
Have you tried the new coding agent in GitHub Copilot? Here’s how developers are using it to work more efficiently.

The post Assigning and completing issues with coding agent in GitHub Copilot appeared first on The GitHub Blog. ⌘ Read more

什麼是 AI 智能體？如何一步步構建屬於你自己的智能體？
什麼是 AI 智能體？下一件大事？Gartner 認爲：AI 智能體（AI Agents）將是未來的關鍵技術。OpenAI、Nvidia 和 Microsoft 正在大力投入，甚至像 Salesforce 這樣在 AI 領域一向低調的公司，也已悄然下注。而毫無疑問，這項技術目前正迅猛崛起。那麼，這一趨勢背後真正的是什麼？理解智能體（Agents）的關鍵在於：自主性（Agency）。不同於傳統的生成 ⌘ Read more

別被炒作迷惑！吳恩達告訴你 AI Agent 智能體的真相與誤區
在 AI 領域，炒作往往走在實踐的前面，今年熱議的 AI Agent（智能體）就是這樣一個充滿誤解與過度承諾的熱門概念。吳恩達（Andrew Ng）無疑是人工智能領域最具影響力的人物之一。作爲斯坦福大學教授、Coursera 聯合創始人、百度前首席科學家、AI Fund 創始人，他不僅推動了深度學習的普及，還培養了一代 AI 人才。另一邊，LangChain 作爲當前最流行的大語言模型應用開發框架 ⌘ Read more

Age of Dead Sea Scrolls pinned down with the help of AI
A new study has used AI to decipher the age of the Dead Sea Scrolls, potentially rewriting what we know about when some of the world’s most precious biblical texts were created. ⌘ Read more

FriendlyElec NanoPi M5 Offers RK3576, Dual LAN, MIPI-CSI, and 6 TOPS NPU
FriendlyElec has introduced the NanoPi M5, a compact single-board computer based on the Rockchip RK3576 processor. It features a 6 TOPS INT8 NPU, supports LPDDR4X or LPDDR5 memory, and offers UFS 2.0 storage along with dual Gigabit Ethernet and MIPI-CSI/DSI interfaces. Compared to recently launched boards such as the NanoPi Zero2 (RK3528A), NanoPi M6 (RK3588S), […] ⌘ Read more

Hack the model: Build AI security skills with the GitHub Secure Code Game
Dive into the novel security challenges AI introduces with the open source game that over 10,000 developers have used to sharpen their skills.

The post Hack the model: Build AI security skills with the GitHub Secure Code Game appeared first on The GitHub Blog. ⌘ Read more

MCP 安全檢查清單：AI ⼯具⽣態系統安全指南
背景本安全檢查清單由 @SlowMistTeam 編寫並維護。慢霧科技作爲全球領先的區塊鏈生態威脅情報公司，旨在幫助區塊鏈生態與 AI 生態相融合的過程中擁有更加深度的安全防護能力及保護用戶擁有更高的隱私性。 感謝 FENZ.AI 傾力相助。FENZ.AI 以未來科技重塑 AI 安全審計。從 MCP 攻擊檢測到 AI 忠誠度評估， Fenz 正在構建 AI 時代的基礎設施——“超級智能，始於 ⌘ Read more

兩分鐘實現一個 chrome 插件
放在以前，2 分鐘寫一個 chrome 插件是想都不敢想，但是今天，通過 ai 賦能，我們可以很容易的就實現了。效果圖—用戶可以設置最大的標籤頁數量和超出數量的管理策略，來管理 chrome 當前窗口的標籤頁。需求場景—-在公司辦公的時候，一個 chrome 頁面，不斷的新開標籤頁，標籤頁越來越多，最後直接都找不到之前想要找到的標籤頁了。所以可以通過這種標籤頁管理插件來管理自己當前的瀏覽器 ⌘ Read more

Paraglider video out of China likely doctored with AI, experts say
The ABC was among hundreds of media outlets which reported on the incredible footage. Experts say the video was at least partially manipulated by artificial intelligence. ⌘ Read more

macOS Tahoe Name Leaked Ahead of Apple’s WWDC Event Next Week
The alleged name of macOS 26 ( yes) has leaked.

Image

In his Power On newsletter today, [Bloomberg’s Mark Gurman said](https://www.bloomberg.com/ … ⌘ Read more

golang 開發 OA 系統，工作流引擎選型
我想開發一套企業內部使用 OA 系統，使用 golang gin 開發，讓 AI 推薦了幾個流程管理的庫，即工作流引擎。發現大部分找到的都是 Java Sprint Boot 的實現，但是我還是希望用 golang gin 去實現。畢竟 golang gin 目前用的最順手，也積累了大量的內部功能庫，而且服務器資源佔用低。加上是集團內幾個公司一塊使用，使用現成的開源實現，也達不到要求，估計也很難改 ⌘ Read more

golang 開發 OA 系統，工作流引擎選型
我想開發一套企業內部使用 OA 系統，使用 golang gin 開發，讓 AI 推薦了幾個流程管理的庫，即工作流引擎。發現大部分找到的都是 Java Sprint Boot 的實現，但是我還是希望用 golang gin 去實現。畢竟 golang gin 目前用的最順手，也積累了大量的內部功能庫，而且服務器資源佔用低。加上是集團內幾個公司一塊使用，使用現成的開源實現，也達不到要求，估計也很難改 ⌘ Read more

MCP Server 的五種主流架構與 Nacos 的選擇
在 AI 大模型應用爆發的今天，Model Context Protocol (MCP) 作爲連接 AI 大模型與應用的關鍵協議，正在快速普及。然而，如何在企業級環境中高效部署和管理 MCP 服務，成爲技術團隊面臨的重要挑戰。本文將深入剖析 MCP Server 的五種主流架構模式，並結合 Nacos 服務治理框架，爲企業級 MCP 部署提供實用指南。MCP 架構的演進與挑戰MCP 協議爲 AI ⌘ Read more

Cursor - 12306 MCP，打造 AI 智能選票系統，超酷的！
大家好，我是蒼何。昨天線下活動見了米尼的孩子，說句實在的，勾起了我對家裏娃的想念。於是，我決定端午，我要回家，我要陪過度六一。掏出手機 12306，查票、候補、買票。開始計算着出發時間，到達家的時間，選擇高鐵還是硬臥，早上 6 點的車能否趕上，晚上十一點的車到家太晚會吵醒孩子，中午的車正好但往往沒票。。。。受夠了，每次都是這樣，買個票，能把我瞬間變成一隻哈巴狗，咧着嘴，大聲喘氣。於是，我想借助 A ⌘ Read more

GitHub for Beginners: Test-driven development (TDD) with GitHub Copilot
See how to use GitHub Copilot to engage in some test-driven development.

The post GitHub for Beginners: Test-driven development (TDD) with GitHub Copilot appeared first on The GitHub Blog. ⌘ Read more

戰鬥系統：事件系統設計
稍有點項目經驗的讀者，想必對事件都不陌生，本篇就和大家分享一下游戲中的事件處理。閱讀提醒：本篇不會講觀察者模式和發佈訂閱模式的基本概念，若尚不瞭解，建議先簡單瞭解一下。 本文字數較多，若想有所得，還需仔細閱讀。 遊戲事件的分類    遊戲中的事件，作者將它們劃爲三類：UI 事件（輸入事件） 養成系統事件（數據變化事件） World 模擬事件（以戰鬥和 AI 爲主） ⌘ Read more

I thought it’s an AI video first. 😆 ⌘ Read more

Top Stories: iPhone 17 Air Details, Apple’s Smart Glasses, and More
WWDC is coming up quickly with a number of software announcements in store, but we’re also looking further ahead to hardware launches like the iPhone 17 lineup and even Apple’s smart glasses project.

Image

This week also saw big news with former Apple design guru Jony Ive joining forces with OpenAI to build future AI-driven devices, while Fortnite return … ⌘ Read more

用 Cherry 替代 Manus，AI 調用多個 MCP 在本地處理 Excel 生成可視化報告文件
Manus 在 agent 領域可謂「炙手可熱」，但受限於網絡以及少得可憐的積分，很多人還是無法用得上。AI Agent 的處理邏輯：無非就是用 AI 根據用戶需求，規劃好要做的事後，不斷的調用不同的工具來實現。有了這個邏輯，我們就知道想要實現某個場景下的需求，只需要找到不同工具的組合給 AI 調用就可以了。而工具方面，各種 MCP 已經成熟了，可以在 AI 對話中直接調用。那豈不是我們自己在本地 ⌘ Read more

Agent mode 101: All about GitHub Copilot’s powerful mode
A full look at agent mode in GitHub Copilot, including what it can do, when to use it, and best practices.

The post Agent mode 101: All about GitHub Copilot’s powerful mode appeared first on The GitHub Blog. ⌘ Read more

@prologic@twtxt.net What I meant, is that I will not say that someone is not really a writer, if they choose to have what they wrote, ran through some spelling and sentence structure checker, like the one included in MS Word, the average phone keyboard, or on reverso.net - given that they look over the output and make sure the corrections make sense.

Similarly, I won’t complain much, if someone uses AI, to remove backgrounds from images, where the AI can preform this task, as well as a human would and makes sure to check it afterwards, or use ai as a way to sort large quantities of images - usually done for science. An example of this, would be having terabytes of plant photos, from some cities camera system and having an AI analyse them, in an attempt to detect notable changes, like mold, parasites, or the plants needing more water.

Google 的 AI 防護方案 - AI Protection
在 Google Next 2025 上，Google 推出一款全新解決方案 -AI Protection，保護整個 AI 生命週期內降低風險。 這是一個結合 AI 與數據安全的方案，非常有前瞻性，值得看看。方案概要AI Protection 通過以下方式幫助團隊全面管理 AI 風險：主要有三個功能：①發現您環境中的 AI 資產並評估其潛在漏洞②通過控制、政策和護欄保護人工智能資產，這是重點③通過 ⌘ Read more

Jony Ive’s AI Product ‘Third Core Device’ After MacBook and iPhone
Details have leaked about the device that OpenAI is developing with former Apple designer Jony Ive. OpenAI is acquiring io, the hardware-based AI startup co-created by Jony Ive, and the first product is one of a family of devices the company intends to launch.

Image

Ive and Ope … ⌘ Read more

[$] LWN.net Weekly Edition for May 22, 2025
Inside this week’s LWN.net Weekly Edition:

• Front: Home Assistant; Setuptools; Debian AI GR; DMA-mapping API; BPF CI; OSPM 2025

• Briefs: Go audit; Oniux; Asahi progress; Rust in FreeBSD; RHEL 10; Rust 1.87.0; RIP John L. Young; Quote; …

• Announcements: Newsletters, conferences, security updates, patches, and more. ⌘ Read more

提示詞注入攻擊的檢測和數據集介紹
提示詞注入攻擊介紹提示詞注入是一種攻擊技術，攻擊者通過精心設計的輸入來操縱 AI 系統，使其偏離原定行爲或繞過設定的安全措施。這類似於軟件開發中的 SQL 注入攻擊，但針對的是 AI 系統的提示詞處理機制。OWASP 把提示詞注入攻擊作爲 2025 年大模型應用風險的第一位，可見其重要程度。相對於原來直接針對大模型的攻擊，在基於大模型的應用中，增加了數據:數據裏也可以加入對大模型的攻擊，舉個例子： ⌘ Read more

ASUS IoT Unveils RUC-1000 Series with 600W GPU Support and Up to 4000 TOPS at Computex 2025
ASUS IoT has announced the RUC-1000 series at Computex 2025, introducing what it describes as the world’s first 2U 19-inch rugged edge AI GPU computer with PCIe 5.0 support for up to 600W GPUs. Designed for edge AI deployments in industrial environments, the new series includes the RUC-1000G and RUC-1000D models, offering performance scalability and […] ⌘ Read more

AMI Aptio V Firmware Powers Radxa Orion O6 at COMPUTEX 2025
AMI has announced that its Aptio V UEFI Firmware will power the Radxa Orion O6 demo platform at COMPUTEX Taipei 2025. Described as the “World’s First Open Source Armv9 Motherboard,” the compact Orion O6 Mini ITX board is designed for AI, edge computing, and multimedia-intensive workloads. The board uses the CiX P1 CD8180 SoC, which […] ⌘ Read more

Again, I was simply pointing out that, if he used AI to correct misspellings, and improve grammar, then this isn’t true:

“This post was written entirely by a human, with no assistance from AI. (Other than spell- and grammar-checking.)”

@thecanine@twtxt.net right. Spell checkers are not AI. Full grammar checking, and correction? That one I have not seeing, but on AI. So, what I meant was, let the grammar gaffes show; we type as we speak (most of the time). About spelling mistakes, well, let them be corrected as we have done since 1971(?).

[$] Debian AI General Resolution withdrawn
Despite careful planning and months of warning, Debian developer Mo
Zhou has acknowledged that the project needs more time to grapple with
the questions around AI models and the Debian Free Software Guidelines
(DFSG). For now, he has withdrawn his proposed General Resolution (GR)
that would have required the original training data for AI models to
be released in order to be considered DFSG-compliant—though the
debates on the topic continue. ⌘ Read more

@bender@twtxt.net @prologic@twtxt.net Jokes aside, I don’t think that’s the right approach either. We had spell checkers, since I can remember, as well as other tools, like the smart image select, used mostly to remove backgrounds. These are tools, that just simplify the process of either opening up a dictionary and looking up a word, you can’t remember the spelling of, or the process of placing a billion little dots around the part of an image you want to select - none of these are creative or enjoyable tasks, we already had tools for them, decades before AI. I don’t think we need to go back to cave paintings, to be free of AIs influence on our creative work.