habr (@habr@zhub.link)
Как я разработал сервис при помощи нейронок, не зная кода
Хотел создать простое приложение, в которое можно будет загнать SEO-трафик и в будущем монетизировать. И при помощи ИИ собрал его примерно за неделю плотной работы, не разбираясь в кодинге. Рассказываю, как это сделал → Привет! Меня зовут Паша Молянов, я руковожу агентством контент-маркетинга и делаю клуб по подписке про нейросети. Бизнесы у меня работают стабильно, поэтому захотелось сделать продукт для себя, буквально на коленке, который в будущем можно будет монетизировать. В статье поделюсь опытом. Я давно слежу за нейронками — ещё в 2016 написал статью о том, как эти штуки будут делать за копирайтеров и редакторов всю их работу. Тогда ИИ был очень глупым, и им развлекались только гики, вычисляя кошачьи мордочки среди фотографий разной ерунды. Всерьёз погрузился в технологии я где-то в 2023, когда появились более-менее адекватные модели ChatGPT и Midjourney. С тех пор начал разбираться с нейросетями, создал «Нейроцех», и теперь использую ИИ каждый день в работе и повседневной жизни.
https://habr.com/ru/articles/944760/
#плагин #chrome #chatgpt #claude #cursor
Show Original Post
tugatech (@tugatech@masto.pt)
DeepSeek prepara agente de IA para desafiar o ChatGPT e o Gemini no final de 2025
🔗 https://tugatech.com.pt/t71510-deepseek-prepara-agente-de-ia-para-desafiar-o-chatgpt-e-o-gemini-no-final-de-2025
#ChatGPT #google #ia #LiDAR #microsoft #mundo #nvidia #OpenAI #software #startup
Show Original Post
jbz (@jbz@indieweb.social)
Man says he called national security officials after ChatGPT sent him into a delusional spiral | CNN
https://www.youtube.com/watch?v=UZdibqP4H_s
Show Original Post
jeresume (@jeresume@piaille.fr)
.....je viens de demander à #chatgpt si l' #Europe est plus névrosée que les #usa ... j'ai aussi demandé à Le Chat et puis à #Apertus #IA si la Suisse est plus névrosée que l'Europe....🤣 encore une fois j'aime bien la réponse de chatgpt et la conclusion ... 😃
Donc, pour répondre simplement :
Non, l’Europe n’est pas plus névrosée que les États-Unis.
Mais elle l’est autrement :
L’Europe rumine.
Les États-Unis fuient en avant.
Ce sont deux versions modernes de sociétés en crise intérieure, mais avec des styles psychologiques et culturels radicalement opposés.
Show Original Post
CuratedHackerNews (@CuratedHackerNews@mastodon.social)
ChatGPT is NOT a LLM – GPT is
https://www.vincirufus.com/posts/chatgpt-is-not-an-llm/
Show Original Post
technews (@technews@eicker.news)
#OpenAI acknowledges that language models like #ChatGPT will always generate false or misleading statements, known as #hallucinations. OpenAI aims to improve their reliability. Additionally, OpenAI advocates for changing #evaluationbenchmarks to reward models for #admittinguncertainty rather than penalising them for it. https://the-decoder.com/openai-says-chatgpt-will-always-make-things-up-but-it-could-get-better-at-admitting-uncertainty/?eicker.news #tech #media #news
Show Original Post
2025 (@2025@warnercrocker.com)
Sunday Morning Reading
A brief breather at home before travels resume, so there’s a full plate for this week’s Sunday Morning Reading including some nostalgia that shouldn’t be, some very interesting reading on AI, a defiant Chicago, and even a bit on gambling and baseball. Enjoy.
Chicago is under threat from a madman and you can feel the tension in the crisp fall air. Dan Sinker has written an excellent Benediction for Chicago On the Eve of Occupation. You don’t hear thoughts and prayers in the check out lines at the grocery store, just a growing sense of defiant preparation.
The pendulum seems to be swinging wildly in the opinion wars about Artificial Intelligence now that some are actually able to sift through the hype bubbles and see what’s what. EmptyWheel has an excellent 4-part series that is more than worth your time. It begins with A Normal Person’s Explainer On What Generative AI Is And Does and continues with The Other Half Of The AI Relationship, Proteins, Factories, And Wicked Solutions, and concludes with LLMs Are Lead.
Follow that excellent series with The Tech Industry Has a Dirty Secret: The More People Learn About AI, The Less They Trust it by Victor Tangermann. For what it’s worth, I’m also seeing AI naysayers riding the pendulum back the other way as they find ways to make some of the tools of this tool work for them. No AI could ever sort this out with a summary.
The Power We Use and The Power We Give is a brilliant piece by Philip Bump. As he transitions from his former job with The Washington Post he’s talking about where choosing to land next and why making the right choice about where to exercise what power the words you use live. This is a complicated moment in history on so many levels, well illustrated in this one man’s piece.
Also, here’s an excellent piece from Bump on the goings on in Chicago called Trump Wants To Make War On Chicago. He Picked The Wrong Fight.
Speaking of complexity, David Todd McCarty wonders why so many men find themselves alone later in life in Boys Don’t Cry, Men Don’t Bond.
Chris Armitage says It’s Time For Americans To Start Talking About “Soft Secession.” I take the point about the term and the concept. I’m not criticizing either or Armitage’s piece when I say this, but hell, when the president of the country mockingly riles up everyone with a threat to declare war on a city, I’m not sure there’s anything “soft” about anything anymore.
NatashaMH tackles political amnesia is what I think is a timely piece worth revisiting more than once. Our capacity to forget or set aside so much, so often, is astounding. Check out Inside The Fortress, Outside The Fire. Here’s the money quote:
As often as I can, I remind them how history is a reminder of the lives we lost and of how stupid we really are. “Senseless to the core. And once we’re done with the bloodshed, we write poetry.”
To clear the palette a bit, check out Tim Newcomb’s piece about how A Remarkable Discovery of A Document Shatters One of Shakespeare’s Biggest Mysteries.
Fact checking may be a dying art, given that most of the world has decided we can each have our own facts. Zach Helfand as a wonderful long piece on The History Of The New Yorker’s Vaunted Fact-Check Department. Too damn bad we have to file this under nostalgia.
Speaking of nostalgia, check out Bettor Up by J.R. Moehringer about gambling and baseball. Yes, it’s about gambling and baseball, but it’s the closest thing I’ve seen to the good sports writing (especially about baseball) that I grew up with.
(Image from Roman Kraft on Unsplash.)
If you’re interested in just what the heck Sunday Morning Reading is all about you can read more about the origins of Sunday Morning Reading here. If you’d like more click on the Sunday Morning Reading link in the category column to check out what’s been shared on Sunday’s past. You can also find more of my writings on Medium at this link, including in the publications Ellemeno and Rome.
#ai #ArtificialIntelligence #chatgpt #technology #Writing
Show Original Post
andrebuilds (@andrebuilds@mastodon.social)
LLM communication pattern collection. Starting with:
"Ah, I see the issue now"
Additional frequent phrases:
"Let me break this down for you"
"Here's a comprehensive solution"
"I apologize for the confusion"
"Let me clarify"
The consistently polite and confident tone of AI assistants creates interesting psychological dynamics in human-AI interaction.
#AI #LLM #ChatGPT #ArtificialIntelligence #TechCommunication
Show Original Post
h4ckernews (@h4ckernews@mastodon.social)
GPT-5 Thinking in ChatGPT (a.k.a. Research Goblin) is shockingly good at search
https://simonw.substack.com/p/gpt-5-thinking-in-chatgpt-aka-research
#HackerNews #GPT5 #ChatGPT #Search #AI #Research #Goblin
Show Original Post
GripNews (@GripNews@mastodon.social)
🌘 : "GPT-5 在 ChatGPT 中搜尋表現驚人（研究妖精）",
➤ : "當 AI 成為你最得力的網路偵探",
✤ https://simonw.substack.com/p/gpt-5-thinking-in-chatgpt-aka-research
: "作者Simon Willison 分享他對整合 GPT-5 的 ChatGPT（他暱稱為「研究妖精」）在網路搜尋方面的出色表現感到驚豔。不同於以往建議使用者不要將聊天機器人視為搜尋引擎，他發現 GPT-5 能夠進行深入且全面的網路搜尋，即使任務看似瑣碎或複雜，也能產出豐富的結果。文章列舉多個實際案例，說明 GPT-5 如何協助他識別建築物、調查商品資訊、釐清歷史事實，甚至深入挖掘建築結構的資料。他特別強調，透過選擇「GPT-5 Thinking」模型，能獲得更詳盡的搜尋結果，並樂於透過語音輸入與其互動，認為這項技術已發展到不需特別標註使用 AI 進行搜尋的程度。",
+ : "太棒了！我一直覺得傳統搜尋引擎越來越難找到想要的資訊，看來是時候試試
#: "人工智慧 #搜尋引擎 #ChatGPT #GPT-5 #網路",
Show Original Post
alessiopomaro (@alessiopomaro@mastodon.uno)
🧠 #OpenAI aggiunge su #ChatGPT la possibilità di creare nuovi "branch" partendo da una chat.
👉 I dettagli: https://www.linkedin.com/posts/alessiopomaro_openai-chatgpt-ai-activity-7370407380616056832-D6cD
___
✉️ 𝗦𝗲 𝘃𝘂𝗼𝗶 𝗿𝗶𝗺𝗮𝗻𝗲𝗿𝗲 𝗮𝗴𝗴𝗶𝗼𝗿𝗻𝗮𝘁𝗼/𝗮 𝘀𝘂 𝗾𝘂𝗲𝘀𝘁𝗲 𝘁𝗲𝗺𝗮𝘁𝗶𝗰𝗵𝗲, 𝗶𝘀𝗰𝗿𝗶𝘃𝗶𝘁𝗶 𝗮𝗹𝗹𝗮 𝗺𝗶𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿: https://bit.ly/newsletter-alessiopomaro
#AI #GenAI #GenerativeAI #IntelligenzaArtificiale #LLM
Show Original Post
pasimako (@pasimako@mastodon.social)
Interacting with ChatGPT sometimes feels like giving a nervous student a multiple-choice question. It picks the answer most likely to be correct, and even when there's no clear answer, it still chooses a random one to avoid losing points. And if I ask, "Are you sure it's A and not B?" it quickly switches to B, even if it's not the right answer.
Show Original Post