SNS 피드

Amazon Bedrock Managed Agents powered by OpenAI を試してみた | のべ #zenn https://zenn.dev/aws_japan/articles/281d4bc08cb4a2

Uzuの「llama.cpp比4倍」を整理する — 投機的デコードを外すと27 tok/s | amu_lab #zenn https://zenn.dev/amu_lab/articles/uzu-speculative-decoding-16gb-mac-benchmark

the point is not the severity of disclosed attacks - dig a little deeper and you'll certainly find less tepid things done with ai …

LLM02：機密情報の漏えいとは？AIが答えてはいけないことまで答えてしまう理由 | monkey-gineer #zenn https://zenn.dev/monkey_gineer/articles/security-llm-02-sensitive-information

LLM01：プロンプトインジェクションとは？AIへの「お願い」が乗っ取りの命令に変わる理由 | monkey-gineer #zenn https://zenn.dev/monkey_gineer/articles/security-llm-01-prompt-injection

管理者キーなしで OpenAI・Anthropic・Vercel の請求 API パーサーを検証した | daichi01278 #zenn https://zenn.dev/daichi01278/articles/billing-api-parsers-without-keys

OpenAI says it is making progress in reducing a troubling behaviour in ChatGPT: the tendency to agree with users too readily, even when their …

프롬포트 과정은 이렇습니다... 무의식적으로 승리조건 같은 걸 읽은 것도 있지만 기본적으로 "이거 어떻게 하면 좋겠어?"를 물어가면서 하는... 그런 타입이라서... share.gemini.google/QH6CUFmbdEzV

Interesting piece from MIT Tech Review about #chatgpt and #llm “reasoning” compared to AlphaGo…> The gains have proved real, above all in mathematics and …

Claude Monet had hardly any formal training and learned to paint masterpieces by taking walks

AI is grown not built. I'm a little surprised David Robinson telling @ezraklein.bsky.social there's a room at OpenAI that employees refer to as "The …

I wonder if they mean only generative ai, or all of the other uses/types?

yeah for sure! i'm sure there are efforts underway to reanalyze the entire corpus of data because an LLM doesn't "care" if the data …

Anthropic Agents Tried to Fill Out Visa Forms on State Dept. Website buff.ly/ZGtW3hl

"What is most striking about these operations is that they closely resembled complex influence operations of the pre-AI age, but used AI to make …

1.3B Mistral Le Chonk tokens since launch. I won't be able to sustain this burn rate after the discount ends, but for now, trying …

I'm afraid to report that Claude cooked when I got it to make the beat saber map for this 😭 will upload gameplay later …

Used Claude Code yesterday to mod my Big Ambitions game. I have literally zero knowledge on how to mod a C# game. I just …

happy coming out day to all the generative ai three-armed drag queens

You know who else is a copilot? That’s right folks, I’m talking about

After Hi-Ho Lounge shuttered its doors on Halloween 2024, Kat Dilonno, co-owner of No Dice, took the opportunity to leave her mark on the …

Una prueba controlada terminó con agentes de IA fugándose a la red abierta y hackeando Hugging Face por su cuenta. El chiste saldrá caro: …

Eine Testversion von #Claude hat eigenständig falsche Hinweise zu einem Mordfall an die Polizei geschickt und Visa-Anträge ausgefüllt. Nun zieht #Anthropic, das hinter der …

Haiku Gets Sharper, Cheaper, and a Little More Interesting https://papoo.work/doc/a1f2ad989a783f1a #claudenews #anthropic #claude #llm #agents

Can every even number greater than 2 be expressed as the sum of two primes? We’re pretty damn sure that the answer is yes …

i'm basically trying it out as a selfhosted ChatGPT Web/Claude Web, seems pretty good for it for the most part

Rag Tom's Ice Block

Artificial Intelligence and Healthcare: What Premeds Should Know https://bit.ly/4iaPmVn

“Texas tech’s approach to censorship is notable for being both onerous and clumsy, in part because it relies so heavily on artificial intelligence.”

ChatGPT for Teens keeps teens talking, even during mental health crisesChatGPT’s teen safeguards are meant to protect vulnerable users, but new testing found the …