Can Demis save Google?

Plus, the top AI books to read in 2024

Author

Steve Brown
April 01, 2024

This Week in AI

Oopsy! In the last issue of Synthetic we promised you an article on how AI will make better beer, but some fumble fingers led to an incorrect link. So the story is back again, and this time the link will work. Our apologies!

Open AI’s release of ChatGPT in November 2022 was a shot across Google’s bows. Google, who invented the transformer technology that ChatGPT is built on, seemed to be caught flat-footed. Within a few months Google announced they would merge their Google Brain internal research lab with Alphabet company, DeepMind to create a unified research organization, Google DeepMind, led by pioneering AI researcher and DeepMind CEO/founder, Demis Hassabis. This article digs into Demis’ background, DeepMind’s storied history, the future of AI, and whether Demis might become the next CEO of Google.

Utah becomes the first state in the USA to introduce AI regulations. The legislation is focused on consumer protection and disclosure, with fines and civil penalties of up to $5,000 for companies who use AI and fail to disclose its use to consumers.

MIT Technology Review interviews Open AI’s Alex Reben, who studied math and robotics at MIT’s Media Lab, about what the new wave of generative models means for the future of human creativity.

Video: The future of an AI-powered workforce

This keynote from SXSW 2024 shares some powerful ideas on the future of work. The first 20 minutes discusses the power of exponentials before getting into the meat of the content. The speaker argues the future of work will be filled with creative generalists who partner closely with AI to vastly expand their skills and capabilities, and that “jobs are dead, but work isn’t” and that roles will be replaced by tasks and projects. Stimulate fresh thinking on the future of work and what it means to you, your company, and humanity with this video (54 mins).

AI Innovation

One of the many research avenues to develop artificial general intelligence is neuromorphic computing, an approach that mimics brain function to build intelligence. This article outlines a neuromorphic approach that uses fluidic memristors (memory resistors) to mimic synapses linking neurons and pass information with ions instead of electrons, potentially paving the way to significant improvements in energy efficiency. The article adds more detail and includes a link to the research paper in Nature.

This important piece of research from Meta didn’t get the attention it deserved when first released, eclipsed by the announcements of Open AI’s Sora and Google’s Gemini 1.5 Pro models. The paper details a new AI approach, joint embedding predictive architectures (JEPAs) that can learn as efficiently as humans. According to Meta, V-JEPA, which is trained on video data, learns in a way similar to that of a baby, learning concepts and the ability to perform tasks with just a few examples. Meta is giving their work away to other researchers in the belief that V-JEPA is an important step in the journey towards AI that’s able to understand the world, plan, reason, predict, and accomplish complex tasks. Click the article for an explainer video and a lot more detail.

When it comes to marketing, what can you use AI for, what are some key AI-powered tools to consider, and how are brand leaders already putting AI to work to have impact on their top and bottom line? This article dives in to answer all these questions, as well as discussing ethical considerations and the future of AI in marketing.

X.ai has announced, but not yet released, the latest version of their Grok model. The benchmarks they share in their announcement are underwhelming but Grok-1.5 has a 128K context window and X claims improvements in understanding and other capabilities over Grok-1.0. Benchmarks and more information in the announcement.

AI Insights

(This time with the correct link!) Craft brewing is all the rage, but making good beer is part art, part science. This article in MIT Technology Review covers how researchers spent five years chemically analyzing over 250 beers and assembling related data on the assessments of human tasting panels and raters. The resulting data set that links chemical data with sensory features was used to train AI models that accurately predict a beer’s taste, smell, mouthfeel, and its likely consumer rating. Click through to the Nature article for extra detail.

This HBR article details recent research on the adoption of AI tools and their conclusions on algorithmic management. They suggest that a more careful rollout of such tools is warranted. A must read for anyone in HR or IT leadership, or anyone interested in the future of work.

This article shares a decent list of books for people wanting to learn more about artificial intelligence. You’ll find a good mix here: technical and non-technical, new and classic titles. If you’re trying to figure out your next read, this is a great starting point.

The next evolution in AI will be a shift from chatbots to autonomous agents, able to work alongside people, offload lower value or boring tasks, and to elevate human performance. We are hurtling towards the era of AI assistants and coaches. This Big Think article explores current attitudes to AI in the coaching community, and how AI might transform the practice.

Toolkit for the Future

A new set of generative AI-powered tools to transform images, remove backgrounds and watermarks, create backgrounds, and visualize products. Deliver one-of-a-kind visual experiences and better engagement on the web. Free to try!

The Manifest AI chatbot helps shoppers find what they need and convert faster, all with a human-like touch. The average customer doubles their add-to-cart and conversion rates and gets 25% higher AOV. Their chatbot also integrates nicely with Shopify and your help desk. Free 14-day trials available.

Notion is a calm, clear place to focus on the things that actually matter to you. Organize information, thoughts, to-do lists, calendars, articles, tasks, notes, and team discussions in one, well-organized space. AI features can create content, translate languages, and let you ask your notes questions. Easy to use, with powerful features and templates to get you started.

Find phone numbers, emails, company org charts, and the people you need to get to. Fill information gaps in your contacts, auto-generate follow up emails, and identify new target markets. Get 50 free credits with no credit card down and discover why over 500,000 companies use Seamless.AI to grow their business.

Invite the HireLogic AI to your next phone, video call, or in-person interview and it will provide specific insights, highlight strengths, skills and potential concerns, and create a full transcript of the interview. And HireLogic is free to try!