{"id":10166,"date":"2022-07-10T17:35:44","date_gmt":"2022-07-10T22:35:44","guid":{"rendered":"http:\/\/blog.jlbn.net\/?p=10166"},"modified":"2022-07-10T17:35:44","modified_gmt":"2022-07-10T22:35:44","slug":"a-wave-of-a-i-experts-left-google-deepmind-and-meta-and-the-race-is-on-to-build-a-new-more-useful-generation-of-digital-assistants","status":"publish","type":"post","link":"http:\/\/blog.jlbn.net\/?p=10166","title":{"rendered":"A wave of A.I. experts left Google, DeepMind, and Meta\u2014and the race is on to build a new, more useful generation of digital assistants"},"content":{"rendered":"\n

Adept AI and other startups are building digital companions that do more than just talk<\/em><\/strong><\/p>\n\n\n\n

Alexa, what\u2019s the future of digital assistants? I don\u2019t how Alexa would answer that question. But looking at the number of top A.I. minds who have recently left big tech companies to create well-funded startups dedicated to building a new-breed of digital assistants aimed at being useful for business, a golden era of digital work companions is likely to be around the corner.

Among this new crop of digital startups is Adept AI Labs<\/a>. The company, which emerged from \u201cstealth mode\u201d earlier this year with $65 million in initial venture capital funding, stands out for its founding team. They include a group of researchers from Google<\/a> Brain who in 2017 invented the A.I. architecture known as \u201cThe Transformer.\u201d<\/a> This algorithmic design has underpinned a huge number of A.I. advances, especially in natural language processing<\/a>, over the past five years. Now, the team that created the Transformer thinks the same basic idea can be used to create more capable, general assistants that will be able to work alongside people to help perform a wide range of business tasks.

\u201cThe problem we\u2019ve carved out is how to get machine to collaborate with humans and actually build things for them,\u201d says Ashish Vaswani, Adept\u2019s co-founder and chief scientist. Vaswani was the lead author on the paper that introduced the Transformer. He says what Adept is building is not simply a better chat bot. \u201cWe want to figure out how to get machines to perform actions for people, not just have conversations with them.\u201d

Vaswani says the software Adept is building will learn through human feedback, not just by ingesting a lot of pre-existing data from text, which is how most large language A.I. systems are trained today. David Luan, Adept\u2019s co-founder and CEO, says that language understanding is a key capability that Adept\u2019s software will have to possess, since language is a major way humans provide feedback. But the system won\u2019t just stop with language. \u201cYou can think of it as a universal teammate,\u201d Luan says. \u201cIf you had another person on your team, what would you shamelessly ask them to do? That\u2019s what we want this software to do.\u201d<\/p>\n\n\n\n

Adept\u2019s first step has been creating software that can follow natural language instructions to perform tasks using other software. In\u00a0a demonstration of this that Adept has posted online<\/a>, its software uses a basic SQL database to perform a variety of tasks. A user types \u201ccan you grab the name and population for every country?\u201d and the software goes ahead and pulls that data from the database and assembles it in a simple table. Then a user asks the software to \u201cmake a bar plot of that,\u201d and the software does so. But the plot is hard to read because it contains too many countries. So the user asks it to just \u201cto show the countries with the 6 highest populations,\u201d and the software comes right back with a much easier to read chart. This time though the labels for the six countries are overlapping, which still isn\u2019t great. So the user types, \u201cGood. But the x axis is still a bit hard to read, can you fix that?\u201d And remarkably, the software does so\u2014by writing the labels on an angle\u2014even though the feedback from the user was not that specific. Later in the demo, the software grabs publicly-available U.S. unemployment figures from the Internet and charts those.

This is what Luan calls teaching the software to \u201cclimb the ladder of abstraction.\u201d Eventually, Vaswani says, he wants the software to be able to take an instruction as abstract and complex as, \u201ctell me how my customers are churning?\u201d and have the software analyze the data and produce a report, all without having to receive additional instructions.

Why didn\u2019t Vaswani and his group just stay at Google and build this general assistant for the tech giant? Well, Niki Parmar, another member of the Google team who left to co-found Adept as its chief technology officer, says that at Google, A.I. research is set up to enhance existing products, not create entirely new product categories. \u201cThis is what excites us about Adept,\u201d she says. \u201cHere we can have both research and product together.\u201d She says Adept plans to have a minimally viable product out with customers within months. \u201cWe are a small team that is very aligned to the mission, and we can move fast,\u201d she says.

In addition to Adept, there are also startups such as\u00a0Cohere AI<\/a>, also founded by veterans of Google Brain, including researchers who worked alongside Vaswani on the Transformer, as well as alumni from Meta\u2019s AI Research division and DeepMind. And there\u2019s Inflection, which was co-founded by former DeepMind co-founder Mustafa Suleyman and LinkedIn co-founder Reid Hoffman. All of these companies are looking to create A.I. to assist humans at a wide variety of tasks.

It will be interesting to watch and see how capable these new digital assistants will really be, which will gain traction and for what uses, and how the major tech companies, such as Google and\u00a0Microsoft<\/a>, will respond to what could turn out to be a formidable threat to parts of their business.<\/p>\n\n\n\n

<\/p>\n","protected":false},"excerpt":{"rendered":"

Adept AI and other startups are building digital companions that do more than just talk Alexa, what\u2019s the future of<\/p>\n","protected":false},"author":1,"featured_media":10167,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[254,3421],"tags":[3552,888],"_links":{"self":[{"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=\/wp\/v2\/posts\/10166"}],"collection":[{"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=10166"}],"version-history":[{"count":1,"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=\/wp\/v2\/posts\/10166\/revisions"}],"predecessor-version":[{"id":10168,"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=\/wp\/v2\/posts\/10166\/revisions\/10168"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=\/wp\/v2\/media\/10167"}],"wp:attachment":[{"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=10166"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=10166"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/blog.jlbn.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=10166"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}