The surge of Large Language Models (LLMs) in 2023 has shown what removing those computing shackles can do. Companies like OpenAI have built bigger and better models and released the best versions to the public. As a result, they’ve reached a tipping point where the AI capabilities are hugely impactful, which has led to a surge of interest and further innovation.  In the cloud space, there are fewer barriers between developing a cutting-edge neural network and then serving it to millions of users.

By many accounts, a machine used to serve ChatGPT-4 cost more than the median home price in the USA this year. Though, to be fair, GPUs are one of the few resources that are scarcer than housing lately. This high cost is feasible mainly due to two things: the same system serves many users, and the bandwidth required for each user (just text or individual images) is very low. Edge applications typically have a much smaller number of users per system (perhaps even just one), and the data is often high bandwidth, like HD video or RF sensors.

With the M1076, Mythic uses analog in-memory computing to break these shackles at the edge and usher in a new renaissance in edge AI computing. The M1076 already enables full-HD AI at <4 watts of power. Built on years of experience, customer feedback, and innovations in hardware and software, Mythic’s M2000 series will bring another order of magnitude leap beyond that in power and cost efficiency. Analog in-memory computing is a new technology, and it will be exciting to see where it can lead us in the coming years.

What attracts Mythic to edge computing is that the latency and bandwidth challenges are so significant and showcase the advantages of Mythic’s technology. The enormous challenges of LLMs are similarly exciting. With the upwards of trillions of parameters required, systems struggle to provide enough memory bandwidth to execute LLMs efficiently. Mythic’s in-memory computing eliminates memory bandwidth problems, so there is a natural fit between these technologies. Our early analysis has indicated order-of-magnitude improvements in each metric of cost, performance, and power over Nvidia and leading LLM startups.

This makes us wonder: what will we unlock if we have those order-of-magnitude improvements in LLM computing? Could we marry high-performance LLMs with edge computing? Google published an exciting on the combination of robotics, computer vision, and LLMs earlier this year. They have also recently shown what a multi-modal LLM could do while watching video. LLMs enable systems to reliably respond to novel situations instead of with only pre-set routines. This makes for fascinating possibilities. If someone has a business full of LLM+CV-enabled robots, they’ll need to do that processing at least on-prem. I am okay with waiting a few seconds for Github Co-pilot to respond with coding suggestions, but edge devices need to respond instantaneously so that today’s robots don’t move the same way jerky 80’s robots did.

Emerging use cases for edge AI are growing. Between unshackling edge computing and the introduction of LLMs, it is impacting the industrial, vehicle, health and safety, AR/VR, RF and signal processing, and intelligent systems industry segments. As 2024 nears and technology requirements adapt to match the pace of society, edge AI allows consumers and designers to better integrate devices into our lifestyles. We will need continued breakthroughs in AI computing to see the full benefits of AI at the edge.

To learn more about the latest AI and edge computing trends and watch the presentation for yourself, visit this webpage.

The post Unshackling Edge AI Performance and LLMs appeared first on Mythic.

Many factories are taking advantage of AI to boost workers’ safety. Computer vision can be used to detect when people are in the path of vehicles or equipment, automatically stopping the machines before an accident occurs. AI is also being used in robotics in places where dangerous or hazardous chemicals are being used, reducing people’s exposure in these environments.

In agriculture, AI is being used to reduce pesticide and fertilizer usage, minimizing the environmental impact. AI is also being used to identify exactly how much water plants need, reducing water usage in a time of increasing droughts. Furthermore, new AI driven robots are helping to pick produce that is difficult for workers to harvest, reducing the amount of strenuous tasks that workers have to do and decreasing their exposure to heat and other harsh conditions.

It’s also worth considering the tremendous potential of AI in the video surveillance industry. Stores are using video cameras with AI capabilities to detect potential incidents before they escalate. These cameras can even detect dangerous conditions like wet floors to prevent accidents. 

Finally, smart cities are adding all sorts of new AI capabilities to improve residents’ quality of life and increase safety measures. For example, AI-powered sensors can measure air quality and water quality and predict future issues. Additionally, many cities are using AI systems for traffic management and directing resources to areas that need it, such as roads that are filled with potholes.

With extremely powerful AI tools at all of our fingertips, we look forward to seeing how the next decade of AI innovation will improve our lives and benefit nearly every industry imaginable.

The post Celebrating AI Appreciation Day appeared first on Mythic.

The post Mythic Raises $13 Million to Bring Its Next-generation Analog Computing Solution to Market appeared first on Mythic.

Mike went into detail about how Mythic’s groundbreaking analog compute-in-memory (CIM) architecture offers an “unfair advantage” with up to 10X lower power and cost in comparison to digital solutions that are becoming increasingly challenging and costly to develop. Mythic’s M1076 AMP (Analog Matrix Processor) impressively delivers up to 25 TOPS in a single chip, with a typical power consumption of 3~4W (compared to up to 30W for a digital processor).

With edge AI applications now deploying Full HD and 4K image sensors, Mike noted the growing need for running high-resolution deep-neural networks at the edge. This was highlighted with a demo of M1076 AMP running 30fps Full HD YOLOv5s6 object detection while only consuming 3.5W of power. By offering very high-resolution AI in a compact, low-power form factor, Mythic will usher in the next generation of applications that take edge AI capabilities to a whole new level.

Mike then previewed the Mythic Gen 2 analog compute architecture, which is expected to offer 6~7X the performance of the company’s first generation. Based on a mainstream, cost-effective 28nm process, Mythic Gen 2 features a scalable chiplet architecture, providing the ability to address a broad range of applications from ultra-lower power wearables to high-performance XR applications.

With its scalable chiplet architecture, along with industry-leading performance and power-efficiency, Mythic Gen 2 will be the ideal solution for a wide variety of future edge AI applications. Check out the video of Mike’s Edge AI Summit presentation here, and stay tuned for further details on Mythic’s products and technology!

The post Mythic Highlights its Power-Efficient AI Solution and Previews Next-Gen Architecture appeared first on Mythic.

Check out the video to learn more about Mythic’s pioneering analog compute-in-memory (CIM) technology, which is driving far greater performance and power-efficiency compared to digital systems. As Mike explains, “this small chip has the same compute power as a large system you’d find in the data center.”

The post How Mythic is Driving the Future of AI appeared first on Mythic.

When Mythic was founded, the writing was on the wall for Moore’s Law as the “really rapidly escalating cost of new process technologies was going to increasingly shut out more and more markets,” Mike explained. Since AI applications are very memory intensive, it was clear to Mike and the founding Mythic team that improvements in architecture alone weren’t going to meet the industry’s growing compute demands.

That’s why Mythic developed its analog computing-in-memory (CIM) approach that allows us to deliver unparalleled performance and efficiency for AI applications – all while using older performance nodes that have benefits for cost and supply chain availability. Mike underscored that “we are beating seven and five nanometer Edge AI chips” with more mature process nodes which offer a “massive cost structure advantage.”

Mike highlighted that Mythic’s solutions are entering mass production this year, as the company targets “really exciting opportunities at the edge, everything from gaming to autonomous systems.” He also explained how “household robotics like Roombas, and lawnmowers … could really benefit from computer vision,” along with drones, robotics, and industrial inspection machines. While Mythic is very “computer vision focused” to capture these fast-growing opportunities, Mike also sees “a lot of new things emerging as well” such as “RF and sensor fusion” which could shape the market in the future.

You can watch the full video here: https://kvgo.com/gs/mythic-june-2022

The post Highlights from Mythic’s Panel on Accelerated Computing with Goldman Sachs appeared first on Mythic.