Vibekollen Nytt

AI-nyheter för nyfikna och byggare.

Modeller, verktyg och händelser som formar AI-världen. Hitta nyheter, video och poddar – och följ vad som händer på Vibekollen.

Följ via RSS

Sök och filtrera

Urval Senaste

1 378 nyheter

Alla nyheter i tidsordning. Använd filtren för att hitta ditt ämne, din podd eller ditt verktyg.

23 september 2026

PoddThe AI Daily Brief

Opus 5.5 vs GPT-6 Sol and Luna

Anthropic and OpenAI release new models on the same day, with Opus 5.5 winning early acclaim while GPT-6 Sol and Luna push the frontier of affordable intelligence. NLW breaks down the benchmarks, real-world use cases, and first reactions—and why model personality, falling costs, and the tools surrounding AI increasingly shape which models people actually use.AIDB Fall Listener Survey - https://aidailybrief.ai/surveyMultiplayer AI Sprint - ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://multiplayerai.ai/⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Brought to you by:KPMG – Research from KPMG and the University of Texas at Austin shows the highest-impact AI users treat AI like a reasoning partner — and those skills can be taught at scale. Learn more at ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://kpmg.com/us/Sophisticated⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Harbor - Invest in the AI ecosystem. ⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠https://www.harborcapital.com/aidaily⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠⁠Hyperagent - Hire a team of always-on agents. New users get $100 in free credits.

23 sep. podcasters.spotify.com

VideoMatthew Berman

GPT-6 Sol and Luna Are HERE!

Tell your agents to use this: https://here.now/r/matthewberman Join My Newsletter for Regular AI Updates 👇🏼 https://forwardfuture.com My Links 🔗 👉🏻 X: https://x.com/matthewberman 👉🏻 Forward Future X: https://x.com/forwardfuture 👉🏻 Instagram: https://www.instagram.com/matthewberman_ai 👉🏻 Discord: https://discord.gg/u7wTTGWhuJ 👉🏻 Spotify: https://open.spotify.com/show/6dBxDwxtHl1hpqHhfoXmy8 Media/Sponsorship Inquiries ✅ https://bit.ly/44TC45V Links: https://openai.com/index/introducing-gpt-6-sol-and-luna/

23 sep. youtube.com

BloggTechCrunch AI

Enveda secures $311M to bring more nature-derived AI drugs into clinical trials

The round valued the AI biotech at $2 billion. It is currently testing drugs that treat skin conditions and preserve weight loss after stopping GLP-1s.

23 sep. techcrunch.com

Claude Code släppte en ny version (v2.1.281)23 sep.

BloggHugging Face

How to Use NVIDIA Warp and MjWarp to Accelerate Robotics Simulation and Learning Workflows

A Blog post by NVIDIA on Hugging Face

23 sep. huggingface.co

VideoAI Engineer

AI Engineer Paris 2026 Main Stage: Google DeepMind, ElevenLabs, Hugging Face & Stripe | Day 2

Live from STATION F, the AI Engineer Paris 2026 Main Stage continues with a full day of technical talks from the teams building and scaling production AI systems. Watch sessions covering model customization, inference infrastructure, voice and realtime AI, coding agents, software factories, and the emerging AI economy. Day 2 main-stage lineup: • Jakob Pörschmann (Black Forest Labs) — From Pixels to Robots • Dorian Lods (ElevenLabs) — Generating Identity at Scale • Harry Mellor (Hugging Face) — How a Transformers Model Loads in vLLM • Arielle Le Bail (Stripe) — Inside the AI Economy: What Stripe’s Data Reveals • Hervé Bredin (pyannoteAI) — Rebuilding a Streaming Model for Voice AI • Charles Frye (Modal) — The Low-Latency Inference Engineering Playbook • Dominic Pajak (Arm) — Eyes Up: Edge AI Innovation • Yannis Psorakis (Cognition) — Agentic Map-Reduce for Large-Scale Coding Tasks • Matt Pocock — Fixing the PR Bottleneck • Maarten Grootendorst (Google DeepMind) — The Efficiency of Gemma 4 • Olivier Teboul (Gradium) — The Missing Layers of Conversational AI This livestream covers the Main Stage. Additional tracks and workshops run concurrently and are not included in this broadcast.

23 sep. youtube.com

BloggTechCrunch AI

ChatGPT mobile app gets voice-based agentic features

Pro and Plus users will be able to use the Work tab on their phones to complete agentic tasks.

23 sep. techcrunch.com

BloggTechCrunch AI

Even Americans who use AI every day are worried about it

The report suggests that greater exposure will not resolve the unease around the technology, nor reduce public support for AI regulation.

23 sep. techcrunch.com

VideoMatt Wolfe

I hate it when they do this!

PSA to all the AI companies: Please stop announcing features like they’re live when they’re still rolling out “over the coming weeks.” This happened for like 4 different demos in my recent AI news roundup video... Is it so hard to just make an announcement when I can actually use the new feature 😅 #ai #ainews #claude

23 sep. youtube.com

VideoAI Engineer

From Scratch to SOTA: Training a 3B State-Space Vision Model — Krishna Prasad Srinivasan, Sarvam

Well under one percent of the Common Crawl corpus that frontier models train on is in an Indian language, even as frontier labs call India a fast growing market. Krishna Prasad Srinivasan says the knowledge exists but was never digitized, and Sarvam's answer is a three billion parameter vision language model, small enough for one GPU, that he says beats document AI models a hundred times larger. The model is unusual twice over. Its backbone is a state space model rather than a transformer, because a page can run to ten thousand visual tokens and quadratic attention gets expensive, while an SSM keeps constant memory. And in late 2025, when most OCR models were monolithic page level VLMs, the team bet on block level OCR wrapped in a layout harness and a reading order harness, which many 2026 releases have since converged on. Training runs as a four stage curriculum: thirteen trillion text tokens across English, 22 Indian languages, math, and code before the model sees a pixel, so a language prior can resolve a smudged word; continual pretraining on three hundred million image text pairs; supervised fine tuning on a hundred million OCR samples; then reinforcement learning.

23 sep. youtube.com

BloggGoogle DeepMind

Advancing Private AI Compute with secure, server-side memory

Introducing private, server-side memory to Private AI Compute for personal AI.

23 sep. deepmind.google

VideoAI Engineer

From VLM/VLA's to Embodied Agents — Armen Aghajanyan, Perceptron AI

Feed a model one hour of video and roughly a million visual tokens go in, yet the loss lands on about two percent of them, since the only ground truth is a transcript or a few labeled frames. Armen Aghajanyan calls that a humongous waste, and predicting every pixel treats a background pixel with the same weight as a gripper tip or a contact point. Perceptron's answer is a perceptive objective that learns which percepts will matter, rather than hardcoding the gripper. The second problem is context bloat from cameras that never switch off. Patch averaging buys ten times compression, but his fix is data sparse mixture of experts, a router that decides per layer which tokens to read and which to skip. Left alone, the model zooms into the graph in a figure and spends more tokens on fruit when asked to segment fruit. Put together, that produced the model his team released a few weeks earlier, trained on a petabyte spanning text, images, video, and trajectories from desktop use to video games, which he says beats a frontier lab's embodied reasoning model at a fraction of the cost.

23 sep. youtube.com

BloggGoogle Gemini

A new wave of Connected Apps is rolling out to Gemini.

Gemini is adding new Connected Apps, including Adobe, Airtable, Linear, Peloton and more to help you easily tackle your to-do list.

23 sep. blog.google

BloggOpenAI

Two years of OpenAI Academy

Marking two years of OpenAI Academy and bringing AI skills to even more communities.

23 sep. openai.com

VideoAI Engineer

You’re Not Thinking Big Enough: Rebuilding Food Systems with AI Agents — Cody Menefee, Firecrawl

Cody Menefee once drove three hours to a processor with ten turkeys strapped to the roof of his Tesla, and reports that a windbreak of birds up top does real damage to your range. He raised them in a Nashville backyard where that is not allowed, nearly quit engineering to farm, and learned how hard it is to make money at it. Instead he asks why only three percent of cattle finish on pasture when it is better for animal, consumer, and land. The answer is labor. Grazing done right means splitting pasture into paddocks with one day of feed and moving the herd, fences, and water every day so the grass rests. GPS collars with virtual fences handle the moving but not the deciding: grass grows differently after drought, rain, or trampling, and today a farmer walks out to look. His proposal is to replace those eyes and drop a language model in the loop. Drones give the best imagery but no jurisdiction allows autonomous flights; satellites are too far away; a trail cam pointed at a measuring stick is the cheap version. Feed the model animal locations, drought conditions, and grass height, and let it suggest the next paddock for a human to confirm.

23 sep. youtube.com

BloggGoogle DeepMind

Gemini 3.8 text-to-speech says hello

Gemini 3.8 Flash-Lite TTS and Gemini 3.8 Flash TTS are our most expressive audio models yet.

23 sep. deepmind.google

BloggTechCrunch AI

YouTube Music gets more conversational with new AI features

YouTube is adding new AI-powered features to YouTube Music that aim to make it easier for listeners to discover music, explore artists, and find podcasts they may otherwise miss. Announced at this year’s Made On YouTube event, the updates include “Ask Music,” a conversational tool for music and podcasts, and “Your Podcast Lineup,” a personalized […]

23 sep. techcrunch.com

BloggGoogle Gemini

Gemini 3.8 text-to-speech says hello

a text card image reading "Introducing Gemini 3.8 Flash TTS and 3.8 Flash-Lite TTS"

23 sep. blog.google

BloggNVIDIA

Sakeena Fiza Helps NVIDIA Hardware Succeed at Scale

When Sakeena Fiza describes her work as a validation engineer at NVIDIA, she does so in terms more befitting a detective story than a world-class engineering lab. “Validation engineers look in the shadows and shine a light into every corner,” Fiza said. “Every time we get a system, our first thought is: how can it break?” And when it does? “I always like to think of it as a mystery to solve,” she said. At NVIDIA, the systems Fiza and her colleagues in the data center systems engineering lab investigate are the engines of the AI era. Her work begins before the rest of the world knows a product exists — in the lab — when a new system first receives power. Components are brought up one by one, boards are integrated, firmware and software teams swarm, and engineers watch for the first signs of life. One of Fiza’s earliest and most enduring memories of working at NVIDIA is the collective joy she experienced when she saw the NVIDIA Rubin GPU working for the first time at a system level. “It literally just said, ‘NVIDIA Corporation Device,’” she recalls.

23 sep. blogs.nvidia.com

VideoAI Engineer

The Best Models Still Reason Like Toddlers — Andrew Dai, Elorian

Show a frontier model part of a chessboard, ask how many white squares are visible, and it answers 32. Andrew Dai's diagnosis: it saw a chessboard, knew chessboards have 32 white squares, and hallucinated the rest. The pattern matching that makes these models superb at naming flowers hurts them once a question needs counting or spatial grounding: they miscount a Catan player's roads from the pieces left off the board, and they miss a robot arm lifting a lid because they cannot hold state across a long video. His test for understanding versus reasoning: if a person can answer in one second, so can the model; if it takes longer, the model falls apart. The benchmarks hide this: one popular reasoning suite uses 32 by 32 pixel images, and a multimodal science exam can mostly be answered without the image. What is missing, he says, is visual thinking. Video generators produce cartoonish explosions because their training data is Hollywood and game engines; detection models are robust but passive.

23 sep. youtube.com

Vibekollen Nytt

Hämtar senaste nytt…