Radar Trends to Watch: June 2022 – O’Reilly


The explosion of large models continues. Several developments are particularly noteworthy. DeepMind’s Gato model is unique in that it is a single model trained for over 600 different tasks; whether or not this is a step towards general intelligence (the ensuing debate may be more important than the model itself), it is an impressive achievement. Imagen by Google Brain creates photorealistic images that are impressive even after seeing what DALL-E 2 can do. And Allen AI’s ara (surely an allusion to the Stochastic Parrots article by Emily Bender and Timnit Gebru) is open source, one-tenth the size of GPT-3, and claims to be more accurate. Facebook/Meta also publishes a large open-source language model, including the model’s training log, which records in detail the work required to train it.

Artificial intelligence

  • Is viewing autonomous vehicles as AI systems rather than robots the next step? A new wave of startups are trying techniques like reinforcement learning to train AVs to drive safely.
  • Generative flow networks could be the next major step in building better AI systems.
  • The ethics of building AI bots that mimic real deceased people seems like an academic matter, until someone does: Using GPT-3, a developer has created a bot based on his deceased fiancée. OpenAI objected, saying building such a bot was a violation of its terms of service.
  • Cortical Labs and other startups are building computers that integrate human neurons. It is claimed that these systems can be trained to perform gaming tasks much faster than traditional AI.
  • Google Brain has built a new text-to-image generator called Imagen that creates photorealistic images. Although the images generated by projects like this are always handpicked, the image quality is impressive; developers claim it is better than DALL-E 2.
  • DeepMind has created a new “general purpose” template called Gato. This is a unique model that can solve many kinds of tasks: playing multiple games, tagging pictures, etc. He sparked a debate if General artificial intelligence is simply a matter of scale.
  • AI in autonomous vehicles can be used to eliminate waiting at traffic lights, increase travel speed, and reduce fuel consumption and carbon emissions. Surprisingly, if only 25% of the vehicles are autonomous, you get 50% of the profit.
  • Macaw is a language model developed by Allen AI (AI2). It is available for free and open-source. The macaw is 1tene the size of GPT-3 and about 10% more accurate in answering questions, although (like GPT-3) it tends to fail questions that require common sense or involve logical tricks.
  • Ai-da is an AI-driven robot that can paint portraits, but is it art? Art is as much about human perception as it is about creation. What social clues lead us to think that a robot is creative?
  • Facebook/Meta created a large language model called OPT, which is similar in size and performance to GPT-3. Use of the template is free for non-commercial work; the code is released as open source, along with documents describing how the model was trained.
  • Alice is a modular and extensible open source virtual assistant (think Alexa) that can work completely offline. It’s private by default, although it can be configured to use Amazon or Google as backups. Alice can identify different users (for which she can develop “likes” or “dislikes”, depending on the interactions).

Programming

  • High-volume event delivery without a message queue: Palo Alto Networks built a system to process terabytes of security events per day without using a message queue, just a NoSQL database.
  • New tools enable workflow management across groups of spreadsheets. Spreadsheets are the original low code; these tools seem to offer spreadsheet users many of the features that software developers get from tools like git.
  • Portainer is a container management tool that lets you mount Docker containers as persistent file systems.
  • NVIDIA has made its Linux device drivers open source. The code is available on GitHub. This is a significant change for a company that has historically avoided open source.
  • A startup named Buoyant builds tools to automate Linkerd management. Linkerd, in turn, is a service mesh that is easier to manage and more appropriate for small and medium businesses than Istio.
  • Are we entering the “third age of JavaScript”? An intriguing article suggests that we are. In this vision of the future, static site generation disappears, incremental rendering and edge routing become more important, and Next.js becomes a dominant platform.
  • Rowy is a low-code programming environment that aims to escape the limitations of Airtable and other low-code collaboration services. The interface looks like a spreadsheet, but it’s based on the Google Cloud Firestore document database.
  • PyScript is a framework for running Python in the browser, mixed with HTML (in some ways a bit like PHP). It’s based on Pyodide (a WASM implementation of Python), integrates well with JavaScript, and may support other languages ​​in the future.

Security

  • Machine learning raises the possibility of undetectable backdoor attacks, malicious attacks that can affect a model’s output but do not measurably detect its performance. Security issues for machine learning are not well understood and do not receive much attention.
  • In a new supply chain attack, two widely used libraries (ctx from Python and PHPass from PHP) were compromised to steal AWS credentials. The attacker now claims that these exploits were “ethical research”, possibly with the aim of earning bounties for reporting exploits.
  • Although not yet accurate enough to work in practice, a new cyberattack detection method can detect and stop attacks in less than a second.
  • The Eternity Project is a new malware-as-a-service organization that offers many different types of tools for data theft, ransomware, and many other exploits. It’s possible that the project itself is a scam, but it appears to be genuine.
  • Palo Alto Networks released research showing that most cloud identity and access management policies are too permissive and 90% of granted permissions are never used. Overly permissive policies are a major vulnerability for cloud users.
  • NIST just released a massive guide to supply chain security. For organizations that cannot digest this 326-page document, they plan to publish a quick start guide.
  • The Passkey standard, supported by Google, Apple, and Microsoft, replaces passwords with other forms of authentication. An app sends an authentication request to the device, which can then respond using any authentication method it supports. Passkey is operating system independent and supports Bluetooth in addition to Internet protocols.
  • Both Google and Mandiant are reporting significant year-over-year increases in the number of 0-day vulnerabilities discovered in 2021.
  • Interesting statistics on ransomware attacks: the ransom is usually only 15% of the total cost of the attack; and on average, the ransom is 2.8% of net turnover (with discounts of up to 25% for prompt payment).
  • Bugs in the most widely used ransomware, including REvil and Conti, can be used to prevent the attacker from encrypting your data.

Web and Web3

VR/AR/Metavers

  • Niantic is building VPS (Visual Positioning System), an augmented reality world map, as part of its Lightship platform. VPS allows games and other AR products to be rooted in the physical world.
  • LivingCities builds a digital twin of the real world as a platform to experience the world in extended reality. This experience includes the history, textures and feelings of a place and, of course, a new kind of social media.
  • New research in haptics enables the creation of realistic virtual textures by measuring how people feel about things. Humans are extremely sensitive to material textures, so creating good textures is important for everything from video games to telesurgery.
  • Google upgrades its search engine for augmented reality: they more fully integrate images into searches, create multimodal searches that incorporate images, text and audio, and generate search results that can be explored through augmented reality.
  • BabylonJS is an open source 3D engine, based on WebGL and WebGPU, developed by Microsoft. This is a strong hint that Microsoft’s version of the Metaverse will be web-based. It will support WebXR.
  • The fediverse is a collection of microblogging social media sites (such as Mastodon) that communicate with each other. Will they become a viable alternative to Elon Musk’s Twitter?
  • Varjo is building a “reality cloud”: a 3D mixed reality streaming service that enables photorealistic “virtual teleportation”. It’s not about weird avatars in a fake 3D world; they record your actions in your real environment.

Material

Design

  • Ethical design starts with redefining success: well-being, equity and sustainability, with good metrics to measure your progress.

quantum computing

  • QICK is a new standardized control plane for quantum devices. The design of the control plane, including the software, is completely open source. Much of the cost of building a quantum device is building the electronics to control it. QICK will significantly reduce the cost of quantum experimentation.
  • The researchers built logic gates using error-corrected quantum bits. This is an important step towards building a useful quantum computer.


Learn faster. Dig deeper. To see further.

Previous Acast launches new audio targeting technology for advertisers
Next Vivaldi 5.3 comes with editable toolbars