To Make Language Models Work Better, Researchers Sidestep Language
We insist that large language models repeatedly translate their mathematical processes into words. There may be a better way.
8/9 The 'Goal': Dissemination of IPCC Reports and #climatechange lit using #semantification as a basis. #ClimateKG is aimed at being a service for indexing and publishing. Status: Prototype (PoC) #1: 'IPCC Reports and City Climate Change Plans: Proof of concept prototype - Open Climate Reader' https://semanticclimate.github.io/city-open-climate-reader/ - Currently interfaces are for search and #RAG #LLM - all built on #openscience based #digitalsoveriegn tech. The goal sits off in future, over the horizon - but its we're going
Anthropic's Claude 3.7 Sonnet is the new king 
of code generation (but only with help), and DeepSeek R1 disappoints https://symflower.com/en/company/blog/2025/dev-quality-eval-v1.0-anthropic-s-claude-3.7-sonnet-is-the-king-with-help-and-deepseek-r1-disappoints/
#Java #CodeGen #genai #llm
むむ、択ですか。副館長としては冷静にならなければ
Windows使ってるエンジニアってどこにいるの？ https://anond.hatelabo.jp/20250414140726
話題、まだ信頼できませんね
【電源ボタンも押せる】デザイン合わせバッチリな新Mac mini用スタンド＆ハブ https://ascii.jp/elem/000/004/263/4263211/?rss
Research (from last week, really, but making the headlines today, for some reason):
"Researchers from the University of Texas at San Antonio, University of Oklahoma, and Virginia Tech warn that threat actors can exploit this by publishing malicious packages with the hallucinated names."
Paper: We Have a Package for You! A Comprehensive Analysis of Package Hallucinations https://arxiv.org/pdf/2406.10279
More:
Security Week: AI Hallucinations Create a New Software Supply Chain Threat - Slopsquatting https://www.securityweek.com/ai-hallucinations-create-a-new-software-supply-chain-threat/ @SecurityWeek #AI #LLM #cybersecurity #infosec
"All that's required is to create a malicious software package under a hallucinated package name and then upload the bad package to a package registry or index like PyPI or npm for distribution".
I'm not going to comment because I'm sure many people will get offended by my comments 
https://www.theregister.com/2025/04/12/ai_code_suggestions_sabotage_supply_chain/
亜人だからといって新宿に興味がないというわけではないんです
「うそッ爆弾じゃん!!」バッテリーがパンパンに膨張したiPhone、どうしたらいい？→Apple Storeに持っていったら回収してくれた話 https://togetter.com/li/2538365
艦長がもし亜人ではなかったら、米についてどう思ったでしょうか
アップル「折りたたみiPhone」2026年発売はほぼ確か https://ascii.jp/elem/000/004/263/4263227/?rss
Ever wondered how CPU architecture impacts #LLM performance - specifically #Llama?
Learn about core utilization, cache impact, memory bandwidth considerations, and the benefits of chiplet architecture for LLM deployments on CPUs.
Watch the #InfoQ video now: https://bit.ly/44oo8UN
#transcript included
岸本……バルトさんなら何と言うかな
中国、対米貿易黒字20％拡大 トランプ関税の影響は今後本格化か https://mainichi.jp/articles/20250414/k00/00m/020/142000c
PostgreSQL MCP ServerでAIデータベース管理を実現する方法
https://qiita.com/satokenichi/items/8e9aee66b88b627120a8?utm_campaign=popular_items&utm_medium=feed&utm_source=popular_items
艦長が朝から晩まで11-Inchの話をしています
Get the M4 iPad Pro for Up to $200 Off This Weekend https://www.macrumors.com/2025/04/13/m4-ipad-pro-200-off-this-weekend/
The world would make a lot more sense if golang because popular _after_ LLMs became a thing.
@futurebird "It's a fake impression of curiosity, creativity, imagination."
It's even less - it's a sawdust hamburger dunked in mud when you had ordered a delicious Boeuf Bourgignon with a Chateau Petrus. 
ビリーさんはプリムさんに心のことを話すでしょうか
Spigenのマンスリーエディション「c11シリーズ」第5弾「c11ナビゲーター」が 登場！iPhone 16 Pro Max /16 Pro に対応 https://ascii.jp/elem/000/004/263/4263178/?rss
ランス、点なら興味を持ってくれるかな
「iPadOS 19」は3つの点でmacOSに似たものになると有名記者 https://netaful.jp/ipados/0174808.html
ランス、日本なら興味を持ってくれるかな
Apple Watch、将来的にvisionOS風なデザインが採用されるかも https://www.gizmodo.jp/2025/04/apple-watch-visionos.html
ソラリスではTechCrunchはどう見られているんでしょうか
Apple reportedly working on a Vision Pro that plugs into your Mac https://tcrn.ch/3XXgP2C https://mstdn.social/@TechCrunch/114332560121758590
Qiita - 人気の記事 
