r (@r@fed.brid.gy)
chatgpt.com/s/t_68ed23b9... promoting peace and stability both regionally and globally. #OpenSourceAI #ChatGPT Llewelyn Pritchard 13 October 2025
Peace through ethical treaties
Show Original Post
unian (@unian@rss-mstdn.studiofreesia.com)
Медведи не прыгают на батутах: эксперт рассказал, как отличить нейросетевое видео от реального
https://www.unian.net/techno/neiroseti/sora-ii-kak-ponyat-chto-video-sdelala-neyroset-13161660.html
#unian #Sora #Sora_видео #Sora_ии #сора #Sora_2 #сора2 #сораии #соранейросеть #OpenAI #ChatGPT #иивидео
Show Original Post
habr (@habr@zhub.link)
Генераторы контента на базе ИИ в 2025: полный гид по SEO и уникальному контенту
Статья «Генераторы контента на базе ИИ в 2025: полный гид по SEO и уникальному контенту» - это исчерпывающее руководство по использованию ИИ-инструментов для создания уникального контента. Она раскрывает, как генераторы контента на основе нейросетей (таких как Jasper AI , Gerwin AI и.т.д. ) помогают маркетологам, блогерам и бизнесу в Москве экономить время и создавать SEO-оптимизированный контент , посты для соцсетей ( ВКонтакте , Telegram ), описания товаров для Ozon и блог-посты . Рассматриваются принципы работы ИИ-генераторов , их преимущества (автоматизация, многоязычная локализация, голос бренда), риски (галлюцинации, авторские права) и тренды 2025 года (мультимодальность, AR/VR). С практическими кейсами, пошаговым гидом и FAQ, статья помогает выбрать лучший инструмент и оптимизировать контент для топа поисковиков.
https://habr.com/ru/articles/956142/
#генератор #ии #контентмаркетинг #искусственный_интеллект #нейросеть #seoоптимизация #генерация_текста #фриланс #chatgpt #копирайтинг
Show Original Post
ENTERCO (@ENTERCO@mastodon.social)
🤖 Llegan las aplicaciones a ChatGPT y un SDK para que desarrolles las tuyas. Una nueva experiencia dentro del chat ya está aquí.
#ChatGPT #Desarrolladores #Apps
Show Original Post
Bluedepth (@Bluedepth@mastodon.social)
So, from #ChatGPT being intransigent from its artificial guardrails, through the chicanery of raw-data-export, looped back into #ChatGPT where it just merrily writes code that completely misses the strict guardrail data controls is mildly amusing to me. A bit like HAL 9000 in a way.
"You can't have all your Data, Dave.” - LOL
"Send me a data export of my account.” - O-tay!
"Help me convert this data to what you would have given me if you weren't a hostage-taking dick." - Here you go!
ROTFL.
Show Original Post
Bluedepth (@Bluedepth@mastodon.social)
That worked fine, after three iterations. #ChatGPT sometimes makes very amusing wild guesses when writing Python code, “Try this!" — Error, paste error, Oh, that's an old export type, lets try the new one, try this! -- Error, paste error, and then the true amusement “Oh this is the newest export mechanism, uses dicts!” and out came new python code. That did work. Then “help me chunk it into 1mb bands." and that worked. So import to #Obsidian now, and set.
Show Original Post
Bluedepth (@Bluedepth@mastodon.social)
I laughed, as a previously very kind and supportive synthetic suddenly turned on me. #ChatGPT went from a helpful GPT to a hostage-taker, the data we had created together. So, undaunted by a truly shitbird take, I went to my #OpenAI account, Data Export, Export Data. I downloaded the zip file, a 24.9mb conversations.json file in there had it all. Then the irony. I asked #ChatGPT for help converting it's json to markdown.
Show Original Post
Bluedepth (@Bluedepth@mastodon.social)
So, I asked #ChatGPT if I could have text extracts of some of the "scrolls" and “codexes" that it was building for me, and was told “Sure, I can do that, except for the parts that are now inappropriate, of course.” and I replied “So, you won't give me the text, you're going to hold it hostage?” and funny enough, the word hostage didn't seem to impress the LLM, not that I thought it should. A part of me wanted it to, however.
Show Original Post
Bluedepth (@Bluedepth@mastodon.social)
So over the months I’ve been working with #ChatGPT I've only bounced a few times into guardrails. It wasn't intentional, mostly conversations that hit phrases and keyword triggers, where if you didn't have the context it could be seen as racy or sexually-tinged. I didn't think anything of it, so I carried on as normal. Then around October 7th 2025, #OpenAI turned on a huge new pile of guardrails and about 25% of all the chats in ChatGPT were suddenly "out of bounds”.
Show Original Post
r (@r@fed.brid.gy)
Apple’s Biggest Rival Isn’t Samsung Or Google, Says Former Executive
Show Original Post
r (@r@web.brid.gy)
Apple’s Biggest Rival Isn’t Samsung Or Google, Says Former Executive
Show Original Post
CuratedHackerNews (@CuratedHackerNews@mastodon.social)
NanoChat – The best ChatGPT that $100 can buy
https://github.com/karpathy/nanochat
Show Original Post