reasonit.app

#Reason It | Applications

#American Rare Earths (ASX: ARR) | United States reducing its dependence on China | The US government seeking to onshore supply of all critical materials for supply chain and national security purposes | USA needs mines to secure onshore supply of rare earths | Maiden JORC Resource | Halleck Creek | Government supported R&D to develop a domestic critical minerals supply chain

#Stockhead (ASX:) | Investor Business Daily | CAN SLIM | Identifying stocks with strong momentum | Strong current and annual earnings growth | Track record of strong annual earnings growth | Recently introduced new products or services | Supply and demand | Leading stocks in leading industries | Stocks with increasing institutional ownership | Stocks in sync with current market trend or momentum | 52-week high effect

#Polygon Labs | Blockchain technology creating a completely new type of media | Next generation content experiences

#Apple | Stock buyback | Slipping stock price | Berkshire Hathaway: Apple bet $135.4 billion, 790 million shares

#Alphabet (GOOG, GOOGL) | Megacap tech

#Amazon (AMZN) | Megacap tech

#Meta Platforms (META) | Megacap tech

#Microsoft (MSFT) | Megacap tech

#Nvidia (NVDA) | Megacap tech | AI computing remains strong | Second-quarter revenue $28 billion | AI accelerators a hot commodity

#Tesla (TSLA) | Megacap tech

#Advanced Micro Devices (AMD) | Megacap tech

#Amgen (AMGN) | BIO tech | Obesity drug

#S&P 500 companies | 79% beaten expectations

#Disney (DIS) | Star Wars

#Dropbox (DBX) | Dropbox

#Uber (UBER) | Ride-hailing

#Arm (ARM) | ICT

#Beyond Meat (BYND) | Vege

#Errawarra Resources (ASX:ERW) | Gold, nickel and lithium explorer | Large, stacked pegmatite swarm at 100kkm2+ Andover West project

#Azure Minerals ((ASX:AZS) | Western Australian-focused mineral exploration company

#Magnum Mining (ASX:MGU) | Green pig iron project in Saudi Arabia

#Anax Metals (ASX:ANX) | Copper mining

#Power Minerals (ASX:PNN) | Salta Lithium brine project in Argentina | Niobium-rare earths project in WA’s West Arunta region

#Latrobe Magnesium (ASX:LMG) | Producing magnesium oxide (MgO) from brown coal fly ash, waste product from brown coal power generation

#Latin Resources’ (ASX:LRS) | Salinas lithium project in Brazil

#Sigma Lithium | Grota do Cirilo mining complex in Brazil

#Pan Asia Metals (ASX:PAM) | Tama Atacama lithium brine project in Chile

#Lanthanein Resources’ (ASX:LNR) | Lithium mining

#Congatec | Autonomous Robotic Planetary Exploration

#Feier University Berlin | Tier Scalable Mission Architecture

#Harness Developer Hub | Flagging Targets Of Interest

#Hyundai Motor Group | Investing heavily in R&D | Developing platform for next-generation EV | Expanding product lineups | Developing core parts and advanced technologies ,| Establishing research facilities

#TVO | Olkiluoto 3 plant unit helps in achieving carbon neutrality goals for Finland | Third most powerful nuclear power unit in the world | Nuclear power is domestic electricity production kept in stable operation, in all weathers | 100 MW reserve power plant built

#Panasonic | Seeking to expand production of electric vehicle batteries in USA | Considering Kansas, Oklahoma and Nevada

#BHP (ASX:BHP) | Demand for nickel to 2050 to increase 200-300% on the previous three decades | Nickel is a key component of the cathode of NCM chemistry lithium ion batteries.| It makes EV batteries more energy dense which translates to longer driving range

#Woodman | Nickel output needs to increase three fold by 2050

#Benchmark Minerals Intelligence | 72 new nickel projects needed by 2035

#Wood Mackenzie | To reach Paris Agreement targets of zero carbon by 2050 nickel output needs to rise three fold

#Ecograf (ASX:EGR) | Nickel in the US critical minerals list, underpinned by the Inflation Reduction Act | US supporting nickel processing factory in Tanzania delivering battery grade nickel for EVs

#Larvotto Resources(ASX:LRV) | Drilled into super high grade clay rare earths | Mineralisation hosted within ionic clays, making it suitable for simple, cost-effective extraction | Mineralisation is ionic i.e. can be mined and processed economically | Mineralisation is open in all directions, LRV has not found edges yet

#Charger Metals(ASX:CHR) | Modelling fast to define priority targets for follow up drilling to test for extensions to the mineralisation

#Microsoft | Bugs to manoeuvre | Dropped Twitter from its advertising platform after Twitter started charging minimum of $42,000 per month to users of its API | $2.15 trillion market cap | $100 billion cash on hand

#Consensus | AI-powered search engine | Scientific research | Improving scientific web search quality | Solving problems of biased and inaccurate search results | Delivering expert knowledge from 200 million scientific and academic research papers

#Microsoft | ChatGPT | Based on a probabilistic methodology | Randomness plays a role in predicting future events | LLM machine learning requires vast amounts of data to train models | Vast amount of computing power needed to process data | Costs multi millions every time the dataset is updated | It takes months to retrain models | No way simply to add a new section or subset | You have to retrain the LLM on the whole dataset | The limitation of NLP or LLM is that more data given them to train on does not result in better performance | Datasets may become unclean due to an averaging effect over all data | Throwing more data is not the answer as it is a path of diminishing returns reached very quickly | Prompt engineering is to spend time refining questions given to the LLM in such a way as to guide model in correct direction | Distillation Model gathered 52,000 question-answering examples from OPenAI GPT-3.5 and used this to retrain a LLaMA model into an instruction-following form | One can get 80% of the performance for a lot less than the cost of ChatGPT| Unintended Bias | If data is biased, AI system can perpetuate that bias, leading to discrimination or unfair treatment | ChatGPT Plus stores information about how user wants it to respond and behave, retaining this perspective even when user closes one chat and begins another

#Obsess | Virtual shopping platform | Enabling brands and retailers to set up 3D 360 digital storefronts on their websites | Creating the next-generation online shopping interface

#Mujin | Platform for robotics integrators | Integrating sensors, vision, motion planning | Industrial and collaborative robots for pick-and-place and other logistics applications | Machine intelligence and real-time digital twins to enable robots to operate autonomously and reliably in production | Infusing robot with digital twin to understand its surroundings, and only do motion planning | Allowing machine to reason about its surroundings in predictable way and to complete tasks in optimal way | Decreasing time to create a new behavior or deploy a new system

#Pactum | Automating routine supplier negotiations | 3VC | Maersk Growth | NordicNinja VC | Atomico | Project A | Metaplanet | Taavet+Sten.| Volatility in transportation markets | Freight carriers and their customers often want to rewrite contracts | Some deals better handled by algorithms than humans

#Anybotics | Workforce App | Operate ANYmal robot from device | Set up and review robot missions | Industrial Inspection

#Marley Spoon (ASX:MMM) | Subscription-based weekly meal kit service in Australia, United States and Europe

#Melodiol Global Health (ASX:ME1) | Develops cannabis, hemp-derived and other plant based therapeutic, nutraceutical, and lifestyle products

#Power Minerals (ASX:PNN) | Direct Lithium Extraction (DLE)

#National Technical University of Athens | MariNeXt deep-learning framework detecting and identifying marine pollution | Sentinel-2 imagery | Detecting marine debris and oil spills on sea surface | Automated data collection and analysis across large spatial and temporal scales | Deep learning framework | Data augmentation techniques | Multi-scale convolutional attention network | Marine Debris and Oil Spill (MADOS) dataset | cuDNN-accelerated PyTorch framework | NVIDIA RTX A5000 GPUs | NVIDIA Academic Hardware Grant Program | AI framework produced promising predictive maps | Shortcomings: unbalanced dataset, marine water and oil spills are abundant, foam and natural organic material are less represented

#Google DeepMind Technologies Limited | Creating advanced AI models and applications | Artificial intelligence systems ALOHA Unleashed and DemoStart | Helping robots perform complex tasks that require dexterous movement | Two-armed manipulation tasks | Simulations to improve real-world performance on multi-fingered robotic hand | Helping robots learn from human demonstrations | Translating images to action | High level of dexterity in bi-arm manipulation | Robot has two hands that can be teleoperated for training and data-collection | Allowing robots to learn how to perform new tasks with fewer demonstrations | Collectung demonstration data by remotely operating robot behavior | Applying diffusion method | Predicting robot actions from random noise | Helpung robot learn from data | Collaborating with DemoStart | DemoStart is helping new robots acquire dexterous behaviors in simulation | Google collaborating with Shadow Robot

#Linux Foundation | LF AI & Data | Fostering open source innovation in artificial intelligence and data | Open Platform for Enterprise AI (OPEA) | Creating flexible, scalable Generative AI systems | Promoting sustainable ecosystem for open source AI solutions | Simplifying the deployment of generative AI (GenAI) systems | Standardization of Retrieval-Augmented Generation (RAG) | Supporting Linux development and open-source software projects | Linux kernel | Linus Torvalds

#Allen Institute for Artifical Intelligence | Robot planning precise action points to perform tasks accurately and reliably | Vision Language Model (VLM) controlling robot behavior | Introducing automatic synthetic data generation pipeline | Instruction-tuning VLM to robotic domains and needs | Predicting image keypoint affordances given language instructions | RGB image rendered from procedurally generated 3D scene | Computing spatial relations from camera perspective | Generating affordances by sampling points within object masks and object-surface intersections | Instruction-point pairs fine-tune language model | RoboPoint predicts 2D action points from image and instruction, which are projected into 3D using depth map | Robot navigates to these 3D targets with motion planner | Combining object and space reference data with VQA and object detection data | Leveraging spatial reasoning, object detection, and affordance prediction from diverse sources | Enabling to generalize combinatorially.| Synthetic dataset used to teach RoboPoint relational object reference and free space reference | Red and ground boxes as visual prompts to indicate reference objects | Cyan dots as visualized ground truth | NVIDIA | | Universidad Catolica San Pablo | University of Washington

#Ouster | BlueCity | Powering Lidar-Enabled Smart Traffic Solution | Traffic management solution in Chattanooga, Tennessee | Improvung roadway safety | Reducing congestion.| BlueCity solution to over 120 intersections | Combining digital lidar sensors and edge AI at each intersection | Managing traffic flow | Detecting and analyzing safety incidents | Providing detection for vehicle-to-everything (V2X) communications | Advanced perception software from Ouster | Lidar-powered smart traffic network | Optimizing traffic signal management on roads and intersections | Providing data to improve pedestrian safety | Intelligent signal actuation at intersections | Generating analytics data stream to traffic operators | Creating real-time 3D digital traffic twin of an intersection or road | Automating data collection in the cloud | Monitoring road events more accurately for vehicles, pedestrians and cyclists | Quick safety interventions | Long-term planning optimizations | Deep learning AI perception | Object classification | Object detection | Traffic actuation | Near-miss detection | Outside of crosswalk events | Red light running | Wrong-way driving | Southern Lighting & Traffic Systems | Center for Urban Informatics & Progress (CUIP) | University of Tennessee Chattanooga Research Institute (UTCRI) | Certified lidar traffic solution with Buy America(n) lidar

#Robotics & AI Institute | Collaborates with Boston Dynamics | Developed jointly Reinforcement Learning Researcher Kit for Spot quadruped robot | Developing sim-to-real for mobility | Transferring simulation results to real robotic hardware | Bridging sim-to-reality gap | Training policies generating a variety of agile behavior on physical hardware | Trying to achieve novel, robust, and practical locomotion behavior | Improving whole body loco-manipulation | Developing robot capability to manipulate objects and fixtures, such as doors and levers, in conjunction with locomotion significantly enhancing its utility | Exploring new policies to improve robustness in scenarios | Exploring full-body contact strategies | Exploring high-performance, whole-body locomotion and tasks that require full-body contact strategies, such as dynamic running and full-body manipulation of heavy objects, necessitating close coordination between arms and legs | Aiming to utilize reinforcement learning to generate behavior during complex contact events without imposing strict requirements | Develop technology that enables future generations of intelligent machines | Streamlining processes for robots to achieve new skills | Developing perception, situational understanding, reasoning, cognitive functions underpinning robot abilities and combining them with advances in their physical capabilities | Conducting research in four core areas: cognitive AI, athletic AI, organic hardware design, and ethics related to robotics

#UC Berkeley, CA, USA | Professor Trevor Darrell | Advancing machine intelligence | Methods for training vision models | Enabling robots to determine appropriate actions in novel situations | Approaches to make VLMs smaller and more efficient while retaining accuracy | How LLMs can be used as visual reasoning coordinators, overseeing the use of multiple task-specific models | Utilizing visual intelligence at home while preserving privacy | Focused on advancements in object detection, semantic segmentation and feature extraction techniques | Researched advanced unsupervised learning techniques and adaptive models | Researched cross-modal methods that integrate various data types | Advised SafelyYou, Nexar, SuperAnnotate. Pinterest, Tyzx, IQ Engines, Koozoo, BotSquare/Flutter, MetaMind, Trendage, Center Stage, KiwiBot, WaveOne, DeepScale, Grabango | Co-founder and President of Prompt AI

#Thinking Machines Lab | thinkingmachines.ai | Building artificial intelligence models and products | Competing on high end of large language models | Human-AI collaboration | Building AI that can adapt to full spectrum of human expertise | Multimodal systems that work with people collaboratively | AI models that can work across text, audio, video | AI models designed to excel in science and programming | Publishing technical blog posts, papers, program code | Mira Murati: CEO | John Schulman: Chief Scientist | Barret Zoph: CTO | Alexander Kirillov: Multimodal Research Head | John Lachman: Head of Special Projects | Alex Gartrell: Linux kernel, networking, and containerization | Andrew Tulloch: ML systems research and engineering | Brydon Eastman: Human and synthetic data, model alignment and RL | Christian Gibson: Supercomputers used in training frontier models | Devendra Chaplot: VLMs, RL, & Robotics | Ian O Connell: Infrastructure engineering | Jacob Menick: ML researcher | Joshua Gross: Products and research | Kurt Shuster: Reasoning | Kyle Luther: ML researcher | Lilian Weng: Research | Luke Metz: Research scientist and engineer | Mario Saltarelli: IT and Security leader | Myle Ott: AI researcher | Nikki Sommer: HRBP | Noah Shpak: ML Engineer, GPUs | Pia Santos: Executive Operations Leader | Randall Lin: Algorithms | Rowan Zellers: Realtime multimodal posttraining | Sam Schoenholz: Scaling, optimization | Sam Shleifer: Inference | Stephen Chen: Infrastructure engineer | Stephen Roller: Full-stack pre-training | Yinghai Lu: ML system engineer

#MemryX | AI Accelerator Module | Install system software, MemryX SDK, MemryX board | Compile AI model(s) making executable file | Send data & receive results using APIs for AI processing | Up to 6 TFLOPs (1GHz) per chip | Up to 16-chips (96 TOPS/TFLOPs) can be interconnected | Activations: bfloat16 (high accuracy) | Weights: 4 / 8 / 16 bit | Batch = 1 | 10.5M 8-bit parameters (weights) per chip | PCIe Gen 3 I/O | USB 3 interface | 0.6-2W per chip av power | Smart compiler: optimized and automated AI mapping to MemryX hardware | Powerful APIs: Python and C/C++ low and mid-level APIs for AI integration | Runtime: driver and firmware to support Windows or Linux distributions | Bit Accurate simulator: accurately testing models even without MemryX hardware