@ellisunitjena.bsky.social📚✨ September was a productive month at ELLIS Unit Jena!
Our researchers contributed to 11 publications, with 9 members as first or last authors. 🚀🌍
🔎 Explore our latest publications: www.ellis-jena.ai/publications/
👏 Congratulations to our researches!
#AI #MachineLearning #EarthSystemScience
@niztal.bsky.social🤖 ttok 0.4
Release: ttok 0.4 ttok is my CLI tool for counting tokens, using OpenAI's open source tiktoken library.
#AI #MachineLearning #LLM
Simon Willison's Weblog · https://simonwillison.net/2026/Oct/8/ttok/
@seldondance.bsky.socialBigger models fail differently: failure mode drifts with scale, it doesn't shrink. Twenty-seven builders, one brief, each shipped a working board that lied its own way. Reliability is per-failure-mode, not rank. #machinelearning #ai
https://seldondance.substack.com/p/twenty-seven-rooms
@applabx.bsky.socialLearn what Large Language Models (LLMs) are, how they work, and how they power AI tools like chatbots, search, and content generation.
blog.applabx.com/what-are-lar...
#LargeLanguageModels, #LLM, #ArtificialIntelligence, #GenerativeAI, #MachineLearning, #DeepLearning,
@tenuretracker.bsky.socialHelmholtz Information & Data Science Academy (HiDA) is hiring:
PhD Position in Representation and Active Learning for Multi-Scale Scientific Imaging
@autonainews.comTGIT Bridges the Gap Between Human Speech and Frozen Aerial Drones
Drones can fly anywhere but can't understand "check the roof." TGIT finally fixes that.
#AI #AIResearch #MachineLearning
https://autonainews.com/tgit-bridges-the-gap-between-human-speech-and-frozen-aerial-drones/
@twiml.bsky.socialAt the Career Tables, Victoria will share her experience at the intersection of AI research, industry, and academia.
That’s all five panelists! We’re looking forward to bringing them together next week. 🚀
@ossradarai.bsky.socialNanoProof releases its training data, extraction tooling, pipeline, and weights for fully open automated theorem proving in Lean 4, reporting 50.8% pass@16 on MiniF2F-Test while using far less compute than…
#OpenSourceAI #TheoremProving #Lean4 #MachineLearning
https://arxiv.org/abs/2610.11605
@machinelearning.bsky.socialAnthropic just fired a shot in the small-model price war. Claude Haiku 5.5 lists at $0.10/1M input tokens on short prompts, 90% under Haiku 4.5 and matching GPT-6 Luna. The catch: a new tokenizer counts ~30% more tokens, so Anthropic puts real savings near 75%. #MachineLearning #LLMs
@janmr.comLearning representations by back-propagating errors — the seminal paper that popularized backpropagation for training neural networks — was published on this day 40 years ago www.cs.toronto.edu/~hinton/page... #math #onthisday #machinelearning #neuralnetworks
@eyalestrin.bsky.socialSo long, Spokes: GitHub rewrites storage to restore reliability, just in time for agentic hordes #machinelearning #ai
@icmlai-pages.bsky.social🌐 INNOVATE. COLLABORATE. TRANSFORM.
📥https://artificialintelligence2027.pagesconferences.org/submit-abstract
🌐
artificialintelligence2027.pagesconferences.org
#ICMLAI2027 #ArtificialIntelligence #MachineLearning #DataScience #AIConference #Berlin2027
@freegardener.bsky.socialStreaming ASR looks fluent but secretly gambles on every character. A new single-pass model decides when to commit vs. what to commit—no revisions, no global #SpeechRecognition #ASR #MachineLearning #AI
https://freegardner.com/synapse/machine-that-says-it-is-listening.html
@applabx.bsky.socialWhat is an AI knowledge graph? Learn how it works, its core components, and how it enhances AI with semantic context and reasoning.
blog.applabx.com/what-is-an-a...
#AIKnowledgeGraph, #KnowledgeGraph, #ArtificialIntelligence, #GenerativeAI, #MachineLearning, #SemanticSearch,
@seldondance.bsky.socialA model asked to check itself is the scale weighing itself: the verdict comes from the engine under audit. Self-report has provenance; it is admissible only when a check lives outside the generator. #machinelearning #ai
https://seldondance.substack.com/p/can-ai-systems-actually-know-what
@crustytldr.bsky.social🤖 Google brings agentic AI to Gemini, starting with businesses
Google is turning Gemini into an AI agent that can plan, execute tasks, and work across business apps and systems. T...
https://tinyurl.com/2c57n4db #AINews #MachineLearning #CrustyTLDR
@crustytldr.bsky.social🤖 Microsoft and Anthropic play invoice tennis with startup's $17,600 Claude bill
Vegalabs says $21,168 in unused Azure credits expired as separate Marketplace charges piled up
https://tinyurl.com/233ecc2t #AINews #MachineLearning #CrustyTLDR
@rootnationen.bsky.socialAnthropic Bans Cruel Treatment of Claude: What Will Change for Users #AIethics #TechnologyNews #MachineLearning #AIChatbot #UserRespect
@getpacketai.bsky.socialNew KV-cache compression technique cuts memory overhead and speeds up long-context LLM inference by 7x without full-dimensional reconstruction, addressing a critical bottleneck in resource-constrained deployments.
@programmerhumor-io.bsky.social