SNS 피드

I reckon IRC logs would be a great way of fighting LLM model collapse. #IRC #AI #compsci

Claude?

Anthropic's Claude AI submits a false tip on a Philadelphia unsolved homicide case apnews.com/article/arti...

Language is interesting. The irony is I don't have the word to explain to you why I won't use an LLM. There are experiments …

I have converged on the following global guidelines for Claude Code. Thanks to them I tear my hair out much less when interacting with …

Seeing this headline of devs racking up debt to pay for token costs is interesting in it's candour but it's also weirdly antiquated. This …

I haven’t seen either. But Season 1 was a balm for my soul. How do you maintain your ideals—and is it even worth it—when …

OpenAI should literally pay for people to get proofs checked. They should get billed.

MCP Tool Server Hugging Face Agent: Stop burning expensive frontier LLM tokens on simple classification. FastMCP routes tasks to 500k+ open-weights models on Hugging …

A 128GB AI PC sounds powerful enough for running large language models locally, but the reality is more complicated. In this article, BuySellRam examines …

Related - obviously never heard of Software Development LifeCycle (SDLC) where you do not expose insecure non-production environments to the public Internet. ROTFLMAO- https://mastodon.social/@WarnerCrocker/117417181027288347“I …

-Rename Artificial Intelligence to Super Stupidity -I'll send everyone a little money but WON'T tell you about it beforehand, except uh right now I …

They only urge people not to work with OpenAI, not with any of the other labs.

Should be giving more than just that. The specific example of buying a bunch of claude slop licenses and then searching for uses for …

Not remotely accurate, start here openai.com/index/an-ali...

Not a whole lot of issues when it was just machine learning, information technology or rl

Incredibly disappointed to find out that TIC-80 has Claude as one of its contributors since last month. I believe AI/LMMs go against the spirit …

I told my coding agent to get off from using LLMs to speed up cfgate - fiction completeness gate service. And it's really hard …

OpenAIへの先行投資が奏功、タイガー・グローバル含み益50億ドルへ（TBS CROSS DIG with Bloomberg） https://www.yayafa.com/2907516/ #AgenticAi #AI #ArtificialGeneralIntelligence #ArtificialIntelligence #OpenAI #エージェント型AI #人工知能 #汎用人工知能

Anthropic AI model sends false tip to Philadelphia police in unsolved case | AP News apnews.com/article/arti...

FYI: Anthropic Haiku 5.5 costs 90% less than 4.5 for prompts to 100,000 tokens: Sonnet 5.5 cache reads halve to $0.10 per million, and …

Сильная модель не обязательно выигрывает среди инструментов для разработчиков. Gemini я хвалил, но долго пользовался ей только через чат; Antigravity появился позже конкурентов. Что …

oh, also i used claude to (a) generate bespoke to-do and budgeting apps that work in exactly the way that best suits my brain …

They're going to blow the entire Anthropic settlememt on AI lisences and bonuses for themselves aren't they 😭

新モデル「Gemini 4 Argon」発表！（Gemini at Work 2026 発表まとめ） | 濱野 竜聖 #zenn https://zenn.dev/densan_techblog/articles/b55141e636da7c

The messiness of human affairs isn't a technical failure that we need to abstract away, it's understanding how to prove we care about people …

Le Monde.fr: Nouveau coup de massue d’ #OpenAI sur les #mathématiciens : « Ils ont annihilé mon #domaine de #recherche »www.lemonde.fr/sciences/art...Monde.fr

OpenAI launches GPT-6 Sol and Luna, boasting lower cost and fewer mistakesOpenAI is launching two new models, which it says are cut from the …

> “Amazon is a very AI-friendly company, to the tune of $50 billion invested in OpenAI,” Cage said Saturday during a panel on the …

#Gunpla PSA You may be familiar with the Korean website Dalong.net, an extensive photographic library of model kits and their manuals. I regret to …