Version à l'épaule du missile LLM (made in Belfast) de Thales Defence
Viseur/lanceur réutilisable avec un canister qui se clipsededans. Et possibilité d'accueillir LLM ou Starstreak sans changement (13 ou 14 kg) pour de la courte portée
Viseur/lanceur réutilisable avec un canister qui se clipsededans. Et possibilité d'accueillir LLM ou Starstreak sans changement (13 ou 14 kg) pour de la courte portée
June 17, 2025 at 10:44 AM
Version à l'épaule du missile LLM (made in Belfast) de Thales Defence
Viseur/lanceur réutilisable avec un canister qui se clipsededans. Et possibilité d'accueillir LLM ou Starstreak sans changement (13 ou 14 kg) pour de la courte portée
Viseur/lanceur réutilisable avec un canister qui se clipsededans. Et possibilité d'accueillir LLM ou Starstreak sans changement (13 ou 14 kg) pour de la courte portée
My nutritionist LLM is working:
1. Have lost 4 kg for now
2. Tightened my belt by one notch
3. Muscle is preserved (in fact, I have more tone)
4. Have a better understanding of what it takes
I measure and target macro %s and caloric restriction daily. Then, plenty of water and physical activity.
1. Have lost 4 kg for now
2. Tightened my belt by one notch
3. Muscle is preserved (in fact, I have more tone)
4. Have a better understanding of what it takes
I measure and target macro %s and caloric restriction daily. Then, plenty of water and physical activity.
January 26, 2025 at 7:47 AM
My nutritionist LLM is working:
1. Have lost 4 kg for now
2. Tightened my belt by one notch
3. Muscle is preserved (in fact, I have more tone)
4. Have a better understanding of what it takes
I measure and target macro %s and caloric restriction daily. Then, plenty of water and physical activity.
1. Have lost 4 kg for now
2. Tightened my belt by one notch
3. Muscle is preserved (in fact, I have more tone)
4. Have a better understanding of what it takes
I measure and target macro %s and caloric restriction daily. Then, plenty of water and physical activity.
A company put an H100 on a satellite and launched it into space. They trained NanoGPT from scratch. A 60kg satellite at ~$3000/kg means $180,000 to train a mostly useless LLM, not including satellite manufacturing costs. So that’s a thing someone did www.cnbc.com/2025/12/10/n...
‘Greetings, earthlings’: Nvidia-backed Starcloud trains first AI model in space as orbital data center race heats up
The company's Starcloud-1 satellite is running Gemma, an open model from Google, marking the first time in history that an LLM has been trained in outer space.
www.cnbc.com
January 6, 2026 at 12:41 AM
A company put an H100 on a satellite and launched it into space. They trained NanoGPT from scratch. A 60kg satellite at ~$3000/kg means $180,000 to train a mostly useless LLM, not including satellite manufacturing costs. So that’s a thing someone did www.cnbc.com/2025/12/10/n...
Learn how to turn your Pandas data frame into a knowledge graph using LLMs. Build your own LLM graph-builder from scratch, implement LLMGraph Rransformer by LangChain, and QA your KG
towardsdatascience.com/building-a-k...
towardsdatascience.com/building-a-k...
Building a Knowledge Graph From Scratch Using LLMs
Turn your Pandas Data Frame into a Knowledge Graph using LLMs. Build your own LLM graph-builder and QA your KG.
towardsdatascience.com
November 26, 2024 at 10:14 PM
Learn how to turn your Pandas data frame into a knowledge graph using LLMs. Build your own LLM graph-builder from scratch, implement LLMGraph Rransformer by LangChain, and QA your KG
towardsdatascience.com/building-a-k...
towardsdatascience.com/building-a-k...
Fully agree. I see LLMs as interpreter between humans and knowledge graphs but the answers should always come from a factual base in a KG and not from the LLM itself
November 3, 2025 at 3:24 AM
Fully agree. I see LLMs as interpreter between humans and knowledge graphs but the answers should always come from a factual base in a KG and not from the LLM itself
Amazing #EKAW24 in Ams. 🚀 Thank you to the organisers for the fantastic conference! 👏
💡 Presented our Scholarly Wikidata paper @nandana-m.bsky.social @SanjuTiwari @FinnArupNielsen @TekRajChhetri @AxelPolleres.
Paper: dx.doi.org/10.1007/978-...
#wikidata #llm #rag #linkeddata #kg #scholarlydata
💡 Presented our Scholarly Wikidata paper @nandana-m.bsky.social @SanjuTiwari @FinnArupNielsen @TekRajChhetri @AxelPolleres.
Paper: dx.doi.org/10.1007/978-...
#wikidata #llm #rag #linkeddata #kg #scholarlydata
December 5, 2024 at 4:35 PM
Amazing #EKAW24 in Ams. 🚀 Thank you to the organisers for the fantastic conference! 👏
💡 Presented our Scholarly Wikidata paper @nandana-m.bsky.social @SanjuTiwari @FinnArupNielsen @TekRajChhetri @AxelPolleres.
Paper: dx.doi.org/10.1007/978-...
#wikidata #llm #rag #linkeddata #kg #scholarlydata
💡 Presented our Scholarly Wikidata paper @nandana-m.bsky.social @SanjuTiwari @FinnArupNielsen @TekRajChhetri @AxelPolleres.
Paper: dx.doi.org/10.1007/978-...
#wikidata #llm #rag #linkeddata #kg #scholarlydata
Recap from #ISWC 2024 in Baltimore, MD: Awesome discussions at the #RAGEKG workshop! 🎉. Thank you to the session chairs @polleres.bsky.social and @fosborne.bsky.social moderating discussions on #KnowledgeGraph + #LLM #RAG systems.
🔗 Papers:
2024.rage-kg.org
#KGs #GenAI #Research
🔗 Papers:
2024.rage-kg.org
#KGs #GenAI #Research
December 5, 2024 at 4:08 PM
Recap from #ISWC 2024 in Baltimore, MD: Awesome discussions at the #RAGEKG workshop! 🎉. Thank you to the session chairs @polleres.bsky.social and @fosborne.bsky.social moderating discussions on #KnowledgeGraph + #LLM #RAG systems.
🔗 Papers:
2024.rage-kg.org
#KGs #GenAI #Research
🔗 Papers:
2024.rage-kg.org
#KGs #GenAI #Research
🎄 PyTerrier Advent 22/25: A more complex pipeline—knowledge-graph–enhanced RAG from our EMNLP 2024 paper TRACE. We build a KG over retrieved docs, then use a transformer to reason over triples for better QA. This pipeline instantiation uses a cache (see 20th advent) on LLM-based KG extraction.
December 22, 2025 at 12:25 PM
🎄 PyTerrier Advent 22/25: A more complex pipeline—knowledge-graph–enhanced RAG from our EMNLP 2024 paper TRACE. We build a KG over retrieved docs, then use a transformer to reason over triples for better QA. This pipeline instantiation uses a cache (see 20th advent) on LLM-based KG extraction.
Hope to see many of you next week in D.C. at KMWorld. We are presenting as part of the co-located Text Analytics Forum (buff.ly/45Zcpcx ) and Enterprise AI World (buff.ly/3MezXmO ) #KG+LLM #KnowledgeGraphs #NeuroSymbolicAI
November 2, 2023 at 12:13 AM
Hope to see many of you next week in D.C. at KMWorld. We are presenting as part of the co-located Text Analytics Forum (buff.ly/45Zcpcx ) and Enterprise AI World (buff.ly/3MezXmO ) #KG+LLM #KnowledgeGraphs #NeuroSymbolicAI
Roky si lámu hlavu s tím, z čeho se vyrábí tyhle městské mozaiky a kde obdobné kostičky sehnat.
(Tohle je před kostelem Cyrila a Metoděje v Karlíně)
LLM napověděl trochu... Ale stejně bez řešení.
1 kg mramorových kostiček není problém, ale křemen... 🤷♂️
(Tohle je před kostelem Cyrila a Metoděje v Karlíně)
LLM napověděl trochu... Ale stejně bez řešení.
1 kg mramorových kostiček není problém, ale křemen... 🤷♂️
July 8, 2026 at 8:34 AM
Roky si lámu hlavu s tím, z čeho se vyrábí tyhle městské mozaiky a kde obdobné kostičky sehnat.
(Tohle je před kostelem Cyrila a Metoděje v Karlíně)
LLM napověděl trochu... Ale stejně bez řešení.
1 kg mramorových kostiček není problém, ale křemen... 🤷♂️
(Tohle je před kostelem Cyrila a Metoděje v Karlíně)
LLM napověděl trochu... Ale stejně bez řešení.
1 kg mramorových kostiček není problém, ale křemen... 🤷♂️
BioMedGraphica: An All-in-One Platform for Biomedical Prior Knowledge and Omic Signaling Graph Generation biorxiv.org/content/10.1...
#AI #KG #LLM #graph
#AI #KG #LLM #graph
BioMedGraphica: An All-in-One Platform for Biomedical Prior Knowledge and Omic Signaling Graph Generation
Artificial intelligence (AI) is revolutionizing scientific discovery because of its super capability, following the neural scaling laws, to integrate and analyze large-scale datasets to mine knowledge...
biorxiv.org
December 10, 2024 at 7:23 PM
BioMedGraphica: An All-in-One Platform for Biomedical Prior Knowledge and Omic Signaling Graph Generation biorxiv.org/content/10.1...
#AI #KG #LLM #graph
#AI #KG #LLM #graph
sift-kg
Turn any collection of documents into a knowledge graph. Extract entities and relationships via LLM, deduplicate with your approval, and explore the result in your browser — all from the CLI.
https://github.com/juanceresa/sift-kg
Turn any collection of documents into a knowledge graph. Extract entities and relationships via LLM, deduplicate with your approval, and explore the result in your browser — all from the CLI.
https://github.com/juanceresa/sift-kg
February 26, 2026 at 8:15 PM
sift-kg
Turn any collection of documents into a knowledge graph. Extract entities and relationships via LLM, deduplicate with your approval, and explore the result in your browser — all from the CLI.
https://github.com/juanceresa/sift-kg
Turn any collection of documents into a knowledge graph. Extract entities and relationships via LLM, deduplicate with your approval, and explore the result in your browser — all from the CLI.
https://github.com/juanceresa/sift-kg
I really like the whoop.com LLM integration. Some of the tone is annoying but this is a very data-rich service and the LLM does a good job breaking down numbers/providing insight. Definitely one of the better use cases I’ve seen.
December 19, 2025 at 9:32 PM
I really like the whoop.com LLM integration. Some of the tone is annoying but this is a very data-rich service and the LLM does a good job breaking down numbers/providing insight. Definitely one of the better use cases I’ve seen.
Tutorial at #SemanticsConf: Making Knowledge Graphs AI-Ready
Boost GraphRAG performance by improving KG quality—spot gaps, fix inconsistencies, and enable accurate, explainable LLM outputs.
#SEMANTiCS2025
Boost GraphRAG performance by improving KG quality—spot gaps, fix inconsistencies, and enable accurate, explainable LLM outputs.
#SEMANTiCS2025
September 3, 2025 at 1:26 PM
Tutorial at #SemanticsConf: Making Knowledge Graphs AI-Ready
Boost GraphRAG performance by improving KG quality—spot gaps, fix inconsistencies, and enable accurate, explainable LLM outputs.
#SEMANTiCS2025
Boost GraphRAG performance by improving KG quality—spot gaps, fix inconsistencies, and enable accurate, explainable LLM outputs.
#SEMANTiCS2025
🚀 Announcing langchain-prolog: Bringing Logic Programming to @langchain.bsky.social .
python.langchain.com/docs/integra...
#LangChain #Prolog #AI #LLM #GenAI #LogicProgramming #SymbolicAI #Ontology #KnowledgeGraphs #KG #SemanticWeb #OWL #RDF #ExpertSystems #AITooling #OpenSource #PythonLibrary
python.langchain.com/docs/integra...
#LangChain #Prolog #AI #LLM #GenAI #LogicProgramming #SymbolicAI #Ontology #KnowledgeGraphs #KG #SemanticWeb #OWL #RDF #ExpertSystems #AITooling #OpenSource #PythonLibrary
SWI-Prolog | 🦜️🔗 LangChain
SWI-Prolog offers a comprehensive free Prolog environment.
python.langchain.com
March 4, 2025 at 10:07 PM
🚀 Announcing langchain-prolog: Bringing Logic Programming to @langchain.bsky.social .
python.langchain.com/docs/integra...
#LangChain #Prolog #AI #LLM #GenAI #LogicProgramming #SymbolicAI #Ontology #KnowledgeGraphs #KG #SemanticWeb #OWL #RDF #ExpertSystems #AITooling #OpenSource #PythonLibrary
python.langchain.com/docs/integra...
#LangChain #Prolog #AI #LLM #GenAI #LogicProgramming #SymbolicAI #Ontology #KnowledgeGraphs #KG #SemanticWeb #OWL #RDF #ExpertSystems #AITooling #OpenSource #PythonLibrary
#Apple Media #KnowledgeGraphs is hiring a Senior #MLE (10+ yrs). Work on real-time KG pipelines, #agentic #AI for verification, and applied #LLM/RAG for Music, Books & Podcasts. Shape how creative content is organized for hundreds of millions.
Apply: jobs.apple.com/en-us/detail...
#Jobs #Hiring
Apply: jobs.apple.com/en-us/detail...
#Jobs #Hiring
jobs.apple.com
November 11, 2025 at 1:54 AM
#Apple Media #KnowledgeGraphs is hiring a Senior #MLE (10+ yrs). Work on real-time KG pipelines, #agentic #AI for verification, and applied #LLM/RAG for Music, Books & Podcasts. Shape how creative content is organized for hundreds of millions.
Apply: jobs.apple.com/en-us/detail...
#Jobs #Hiring
Apply: jobs.apple.com/en-us/detail...
#Jobs #Hiring
Two papers on agentic retrieval at EMNLP 2026: SPLIT-RAG partitions KGs for selective multi-agent retrieval on complex queries, delivering top KG+LLM results and lowest latency (29-34s). RELOOP unifies text/table/KG evidence via HSEQ iteration for highest multi-hop accuracy with 3-4 steps.
September 22, 2026 at 4:09 PM
Two papers on agentic retrieval at EMNLP 2026: SPLIT-RAG partitions KGs for selective multi-agent retrieval on complex queries, delivering top KG+LLM results and lowest latency (29-34s). RELOOP unifies text/table/KG evidence via HSEQ iteration for highest multi-hop accuracy with 3-4 steps.
O solucionismo do tech bro em todo o seu esplendor! Cada resposta parece ter saída do gerador de lero lero. Ele fala como se Big Tech fosse uma organização, não a produção tecnológica no capitalismo tardio. Vamo ali na feira comprar um kg de "LLM com fogo" do pequeno produtor, então, gente!
December 18, 2024 at 4:51 PM
O solucionismo do tech bro em todo o seu esplendor! Cada resposta parece ter saída do gerador de lero lero. Ele fala como se Big Tech fosse uma organização, não a produção tecnológica no capitalismo tardio. Vamo ali na feira comprar um kg de "LLM com fogo" do pequeno produtor, então, gente!
Interesting post on Kurt Cagle’s Substack “The Ontologist” open.substack.com/pub/ontologi.... Gives equivalent of our “AI sandwich” (osf.io/8v5bd) for how people now use LLMs with knowledge graphs - LLM translates human’s English question for KG, then translates KG's answer into English for human.
November 17, 2025 at 6:13 AM
Interesting post on Kurt Cagle’s Substack “The Ontologist” open.substack.com/pub/ontologi.... Gives equivalent of our “AI sandwich” (osf.io/8v5bd) for how people now use LLMs with knowledge graphs - LLM translates human’s English question for KG, then translates KG's answer into English for human.
Groß mit Waffe im Anschlag durch die Weltgeschichte eiern ist da nicht.
Das G95 ist (Soweit ich das verstehe) auch nur eine verbesserte Variante der MR223/HK416 und damit ca. 1 Kg leichter. Jetzt ersetzen wir das Glas durch ein Aimpoint, Schalli, Magazin, Foregrip, LLM, Munition... da geht was.
Das G95 ist (Soweit ich das verstehe) auch nur eine verbesserte Variante der MR223/HK416 und damit ca. 1 Kg leichter. Jetzt ersetzen wir das Glas durch ein Aimpoint, Schalli, Magazin, Foregrip, LLM, Munition... da geht was.
May 20, 2025 at 9:14 AM
Groß mit Waffe im Anschlag durch die Weltgeschichte eiern ist da nicht.
Das G95 ist (Soweit ich das verstehe) auch nur eine verbesserte Variante der MR223/HK416 und damit ca. 1 Kg leichter. Jetzt ersetzen wir das Glas durch ein Aimpoint, Schalli, Magazin, Foregrip, LLM, Munition... da geht was.
Das G95 ist (Soweit ich das verstehe) auch nur eine verbesserte Variante der MR223/HK416 und damit ca. 1 Kg leichter. Jetzt ersetzen wir das Glas durch ein Aimpoint, Schalli, Magazin, Foregrip, LLM, Munition... da geht was.
The question posed here is whether it is sufficient for an LLM to be trained to be "honest" wrt that decision process and accurately share the content of its learned KG, or if what we really care about is that the LLM learns an accurate KG for that decision process to operate over.
June 19, 2025 at 5:07 PM
The question posed here is whether it is sufficient for an LLM to be trained to be "honest" wrt that decision process and accurately share the content of its learned KG, or if what we really care about is that the LLM learns an accurate KG for that decision process to operate over.
I also think KG is the way of the future. Because the cost of insertion is small the LLM can form new memories (KG associations) from user interactions in pretty much real time. I'm tired of tabula rasa agents.
February 27, 2025 at 3:34 PM
I also think KG is the way of the future. Because the cost of insertion is small the LLM can form new memories (KG associations) from user interactions in pretty much real time. I'm tired of tabula rasa agents.
Extracting #KnowledgeGraphs (great for #GenAI context) from large docs using a local #LLM. Also re Onya (open source KG tool), treating the graph durably, with network #analytics fully reincorporated into the store.
#Python #LLMOps #AI
loom.ogbuji.net/tech/buildin...
#Python #LLMOps #AI
loom.ogbuji.net/tech/buildin...
Building and working knowledge graphs with Onya and local LLMs
Knowledge graphs are powerful tools for GenAi context. Learn how to work with them using a small, private/local model, building knowledge structure from large documents, and treating the graph as a fu...
loom.ogbuji.net
July 30, 2026 at 4:25 PM
Extracting #KnowledgeGraphs (great for #GenAI context) from large docs using a local #LLM. Also re Onya (open source KG tool), treating the graph durably, with network #analytics fully reincorporated into the store.
#Python #LLMOps #AI
loom.ogbuji.net/tech/buildin...
#Python #LLMOps #AI
loom.ogbuji.net/tech/buildin...
napkin math: a 50 km commute + a hamburger for lunch is about 14 kg CO2 per day
the jury is out on what we're using for CO2 per LLM query, but let's say it's high: 5 g (or sub in your own preference): that's 2800 queries per day
I don't even use that many per week
the jury is out on what we're using for CO2 per LLM query, but let's say it's high: 5 g (or sub in your own preference): that's 2800 queries per day
I don't even use that many per week
August 30, 2025 at 2:36 PM
napkin math: a 50 km commute + a hamburger for lunch is about 14 kg CO2 per day
the jury is out on what we're using for CO2 per LLM query, but let's say it's high: 5 g (or sub in your own preference): that's 2800 queries per day
I don't even use that many per week
the jury is out on what we're using for CO2 per LLM query, but let's say it's high: 5 g (or sub in your own preference): that's 2800 queries per day
I don't even use that many per week