Sick of reviewers asking for unreasonable or costly experiments? Scott Aaronson asked for more results which cost **$400,000**, for a new paper that demonstrates quantum computational advantage on the task of BosonSampling. scottaaronson.com/blog/?p=5122

💥Today, LightOn unveils its Appliance, the world's first commercially available photonic co-processor for AI and HPC Pre-order yours now at lighton.ai/lighton-applia… Leasing starts at 1900€/mth or about US$2250/mth Press release: lighton.ai/lighton-applia… The future is now!

💥🎉LightOn lance PAGnol, le plus grand modèle IA de langue française Annonce presse: lighton.ai/news/ Amusez-vous avec PAGnol sur le site de démo: ⤵️ pagnol.lighton.ai Site technique: lair.lighton.ai/pagnol/

💥This is exciting! Today, LightOn unveils its plan to join the quantum computing space with the LightOn Qore series of quantum photonic processors. For more, read our press release ⬇️⬇️⬇️ lighton.ai/news/

🚨Wanna know more about LightOn's future quantum product? Here is the blog post for you: LightOn Qore in a Nutshell lighton.ai/lighton-qore-i…

💥💥 Today, we are announcing Muse! Here is the press release: “Create, Process, Understand, and Learn with LightOn Muse, A simple API to bring human-like understanding & skills to any application, in any language” The full text: lighton.ai/news/

🎉Start working on our photonic hardware faster, as we just migrated our LightOn Cloud infrastructure to #containers. Blog post by our Project Lead Charles Brossollet "Bringing Containers to light" lighton.ai/blog/bringing-… #developers #Linux #Docker

Why is this important? Quality improves! Local knowledge and language variants are better captured than by translation. #NVIDIA #GTC22

🌍 The next big step for NLP is to bring the power of very large language models to many more languages beyond English. ⬇️ A short thread, to go along with my #GTC22 talk on NLP beyond English with NVIDIA's Megatron: reg.rainfocus.com/flow/nvidia/gt…

🎇#MuseAPI, the first AI fully native in five languages, will launch in just a few hours!🎇🥳👍 Join us tomorrow on 👉muse.lighton.ai, and explore the power of LightOn 's large language models VLM-4!

🎇🍾Today is the day: LightOn officially launches #MuseAPI in public access! Muse is the first AI that speaks natively five European languages🇪🇺 Discover the perfect AI assistant for your text-related tasks at 👉muse.lighton.ai

New preprint out on ArXiv today! Very proud of our experimental study of the capabilities of the Qore photonic processors for the implementation of large circuits with up to 8 input and 38 output modes : arxiv.org/abs/2205.01704 #lighton #quantumcomputers #quantumphotonics

📺Notre Directeur Général et co-fondateur, Laurent Daudet était l'invité de François Sorel dans l'émission Tech & Co de BFM Business pour parler de Muse (muse.lighton.ai). Pour le replay 👇 bfmtv.com/economie/repla…

It was super fun to sit down with Iacopo from LightOn to discuss the Muse API and all things AI. Great to learn about the importance of native language models and get some of our community questions answered - youtube.com/watch?v=B1hanH…. #ai #llms #artificialintelligence #llm

[Today at 4 PM] LightOn will be featured on France Culture . Get ready to learn about the cutting-edge technology and innovation that LightOn is bringing to the table. Don't miss out on this fascinating discussion. bit.ly/40Ti42C #LightOn #ArtificialIntelligence

🚀Supercharge your business with Super Summary - the ultimate tool for unlocking information's power! Easily distill large volumes of docs into actionable insights, enabling informed decisions. Trust our unyielding commitment to data security. Elevate your business today!

Si l’IA générative et ses applications au sein des entreprises vous intéressent, #RDV ce soir avec LightOn invité de l’émission Tech&Co sur BFM Business vers 21h30.

🦉Turn Chaos into Clarity, fast! Introducing LightOnOCR-1B, a SOTA 1B parameter vision language model for OCR that pushes the Pareto frontier. With LightOnOCR-1B, every contract, report, and archive becomes living, searchable intelligence, instantly. 🔹 5× faster than

Thrilled to finally share our new OCR model, built with staghado and Baptiste Aubertin! A compact, blazing fast, and high quality end-to-end parser — we think it’s a great base for the community to build on. Read the blog post for benchmarks, technical details & finetuning demos 👇