AI-WATCH NEWSLETTER VOLUME 23

January 28th, 2024

Happy New Year! HERE WE GO!

Please Contribute Your AI-Related Content:

If you send content before January 14, it will be considered for inclusion in this issue.

Click on gif to see full video.

Our cover gif is from the Stereo 3d Short of Zoo Lights, the holiday animatronic light show at the Santa Barbara Zoo shot with our own hardware designed and built by us, and edited by Christina Love. Please see the full video below in the Daily Life section of this volume.

 

By Gordon Rogers, Gary Lapman, Mark Miller, Richard Bradford 

AI-WATCH 

Courtesy of worldview4d.com worldview4d.com/ai-watch 

A monthly newsletter and blog focused on our opinions of the most significant developments in artificial intelligence and their relevance to the real world. Releases the last Sunday of the month at Noon, PST. 

 

Contents: This week’s focus: Optical Communications

Tickler of the Week: Arg

Note from worldview4d.com: NEW WORLD, BS

Daily Life: VR Animatronic Light show (Optical Communications?); Deep faked Taylor Swift

Government Activities: NOAA; The Five Hubs; Growth, Cozy Bear bites again!

Business News: Just What We Need, an AI Rabbit;

General AI: Tech Analysis of LLM on general reasoning

New Business: Optics in AI and Quantum Computing: 

Conclusion: Spherical Arguments


Tickler of the Week:

EVIL TWIN?
Human: “You cool?”
QAI: “To quote Popeye, the sailor man, AI’ am what eye am”

 

Note from worldview4d.com:

In the coming year we can anticipate a flood of crucial technology convergences: AI AND politics, climate, environment, geoengineering, transportation, military, economics, means of production, insurance, housing, immigration, labor, food, energy, water, waste, health care, general medicine, disinformation, psychology, fitness, education, neuroscience, genetics, virology, and even physics and spirituality, for better AND for worse-

Our effort here is to detail our most significant findings in these individual subjects and more as we find them, and to elucidate the potential relationships between them. Other than these humble undertakings, we are totally bored-

 

Let’s start with:

last weeks overview of geopolitics and disinformation-
“Even the illusion through the use of AI tools that the integrity of an election cannot be trusted … can potentially cause a lot of chaos”, Co-Managing Director David Salvo said on a panel about foreign interference hosted by the Canada School of Public Service.

 

Securing Democracy Dispatch
January 18, 2024
Announcements
ASD is teaming up with GMF Geostrategy for a GMF 2024 US Elections newsletter! This monthly newsletter will provide regular analysis of the elections, with a focus on GMF experts’ insights into threats to the US elections and analysis of transatlantic policy issues. Sign up here!
Our Takes

“Outpacing China no longer simply means outrunning the [People’s Republic of China], but also impeding its capabilities in select game-changer and ‘force-multiplier’ technologies”, Senior Fellow Lindsay Gorman told the Senate Committee on Banking, Housing, and Urban Affairs in her testimony on US-China competition in emerging technology.

The US Congress can play a constructive role in shaping a tougher US approach to transnational repression efforts, Co-Managing Director David Salvo and Research Analyst Nathan Kohlenberg write in testimony to the House Committee on Homeland Security Subcommittee on Counterterrorism, Law Enforcement, and Intelligence.

“Even the illusion through the use of AI tools that the integrity of an election cannot be trusted … can potentially cause a lot of chaos”, Co-Managing Director David Salvo said on a panel about foreign interference hosted by the Canada School of Public Service.

Hamilton 2.0 Analysis

Russian diplomats and state media focused on two main narratives this week:

  • Red Sea attacks: Russian state media devoted significant space last week to the US and UK-led airstrikes against the Houthis. “Red Sea” and “Houthis” were the first and second-most used key phrases in articles published on monitored state media websites. The dominant narratives were that the strikes represent yet another failure of US foreign policy in the Middle East, that they reveal fissures within the United States and among Western allies, and that they will ultimately be unsuccessful.
  • ​“Fraud” in US elections: Last week, multiple state media outlets amplified Russian President Vladimir Putin’s claim that the 2020 US presidential election was “falsified through voting by mail”. Putin’s allegation that Democrats “bought ballots for $10” was in response to a question about voting in Russian occupied regions of Ukraine, which Putin claimed represented true democracy because people “voted with their feet”.

Chinese diplomats and state media focused on two main narratives this week:

  • Taiwan’s election: Chinese messaging zeroed in on the Taiwanese election results, which CGTN declared “reveal that the DPP [Democratic Progressive Party] cannot represent the mainstream public opinion on the island”. Chinese diplomats and state media celebrated Nauru’s switch of diplomatic recognition from Taiwan to China. And,

https://email.gmfus.org/cv/924b95184fcc83e0904bc923e694f288fb69969d

Daily Life:

How I Did It

 

 

https://www.hollywoodreporter.com/news/general-news/taylor-swift-ai-nudes-twitter-x-1235806979/

Government Activities:

“Microsoft says Russian state-sponsored hackers spied on its executives

Jan 19 (Reuters) – Microsoft (MSFT.O), said on Friday that a Russian state-sponsored group hacked into its corporate systems on Jan. 12 and stole some emails and documents from staff accounts.

Midnight Blizzard is also known as APT29, Nobelium or Cozy Bear by cybersecurity researchers and linked to Russia’s SVR spy agency, according to U.S. officials. The group is best known for its intrusions of the Democratic National Committee surrounding the 2016 U.S. election.”

https://www.reuters.com/technology/cybersecurity/microsoft-says-it-was-hacked-by-russian-state-sponsored-group-2024-01-19/

Resources

https://www.noaa.gov/ai/resources

Securing Democracy Dispatch
January 25, 2024
Announcements
ASD is teaming up with GMF Geostrategy for a GMF 2024 US Elections monthly digest to provide analysis of the elections, with a focus on GMF experts’ insights into threats to US elections and analysis of transatlantic policy issues. Sign up here!
Our Takes

Although Americans may be focused on domestic challenges, foreign actors’ attempts to compromise the vote in the 2022 US midterm elections remind us that foreign interference remains a pressing concern for the 2024 elections and beyond. Americans should remain vigilant, Co-Managing Directors David Salvo and Rachael Dean Wilson write in Defense One.

The People’s Republic of China (PRC) is utilizing its state broadcaster CGTN to attempt to convince African audiences that Western democracies are ineffective and active hindrances to the continent’s development, Research Analyst Etienne Soula finds in ASD’s recent analysis.

The suspicious robocall featuring what sounded like President Joe Biden’s voice telling New Hampshire residents not to vote featured serious “tipoffs” of audio clips generated by artificial intelligence (AI), such as an “unnatural, robotic” cadence, Senior Fellow Lindsay Gorman told NBC News.

Hamilton 2.0 Analysis

Russian diplomats and state media focused on two main narratives this week:

  • Sweden in NATO: Despite Russia’s continued insistence last week that NATO is an organization focused on destroying Russia and that NATO expansion triggered the war in Ukraine, Türkiye’s decision to approve Sweden’s accession to NATO was met with relative disinterest from Russian media and diplomats. Kremlin spokesperson Dmitry Peskov called the vote Türkiye’s “sovereign decision”, and state media covered the news dispassionately, with even the usually bellicose TV presenter Vladimir Soloviev saying that the move only caused “slight bewilderment in our Fatherland”.
  • Trump and Ukraine: Russia’s top diplomat cast doubt on former US President Donald Trump’s promise to end the war in Ukraine “within days”, with Sergey Lavrov stating that Kyiv is “not ready for any resolution” and that a Trump presidency “won’t change anything”. At the same time, Russian state media amplified former UK Prime Minister Boris Johnson’s endorsement of Trump, promoted Trump’s business successes, and cited Trump’s praise of Russian President Vladimir Putin, Chinese President Xi Jinping, and Hungarian Prime Minister Viktor Orbán.

https://mail.google.com/mail/u/1/#inbox/FMfcgzGwJvfRDSnMHdlNZVXTjSCQTQjz

Business News:

“Microsoft will pitch in $20 million in cloud computing credits for a U.S. government pilot program to democratize access to AI tools for scientists, including researchers, physicists and biologists.”

https://www.semafor.com/article/01/24/2024/silicon-valley-teams-up-with-the-us-government-on-ai

Rabbit’s r1 AI device took CES 2024 by storm — what is it and why might you want one

By Ryan Morrison

More than 20,000 pre-orders in 48 hours

(Image credit: Future)

Artificial intelligence startup Rabbit unveiled its r1 AI device at CES 2024, quickly becoming a show favorite and within 48 sold out two runs of 10,000 units. It also won the Tom’s Guide CES 2024 best in show for being a “revolution in actionable AI.”

The retro orange device was imagined by the famed industrial design house Teenage Engineering for Rabbit and sits firmly in the new category of AI-first hardware.

The r1 is a smart agent, able to understand natural language requests thanks to the combination of a ChatGPT-style large language model voice assistant and a breakthrough innovation called the Large Action Model (LAM).

Priced at $199 and without any ongoing subscription charges, it has quickly become one of the frontrunners for the must-have gadget of the year. But is it worth the hype?

What is the Rabbit r1?

The Rabbit r1 acts like a personal assistant, taking some of what you might expect to be possible with Siri or Google Assistant and putting it into a cute little orange box.

Unlike the voice assistants in your smartphone, it can take complex requests, break it down into different tasks and perform those tasks on your behalf. It removes the need for individual apps.

For more than a decade we’ve started to see standalone hardware devices being replaced with apps and software services. The idea being that you just need one tool to rule them all, but Rabbit found and CEO Jesse Lyu believes this has actually made things more complicated.

It can do x amount of things that your phone can do, but MUCH faster and more intuitive.

Jesse Lyu, Rabbit CEO

His idea was to create a new type of device that listens to your needs and does everything required to complete the task. Lyu says a standalone device allows them full control over the user experience and an easy upgrade path with new abilities.

It features an LCD display, support for 4G LTE and WiFi connectivity, a push-to-talk button that makes it feel like a walkie-talkie and a rotating camera for front and rear image capture.

The r1 can use audio and visual inputs with its AI models. One example given during the keynote at CES 2024 was to show the device a spreadsheet and ask it to add a new column with specific data — it went off, did the work and emailed an updated version of the entire sheet.

How does the Rabbit r1 work?

The Rabbit r1 has a dedicated operating system called Rabbit OS that utilizes the large action model to call agents known as “rabbits”. Each rabbit can carry out a different assigned task to complete the request from the user.

For example, if you want an Uber to get six people to the restaurant by 6pm a rabbit will hook into the Uber API, look for options, and present the choice back to the large language model. It will then present that to the user who can then press a button on the screen to confirm the ride.

Once the user confirms another rabbit will go back to Uber, send all the relevant information, pre-submitted payment information and return with confirmation and tracking information.

What can it do I can’t do on a smartphone?

Doing the same thing on a smartphone involves finding the Uber app, entering the location, clicking the required vehicle size, adding payment information and clicking confirm.

Rabbit does it all for you from a voice prompt. Lyu says the goal is to reduce the amount of clicking required to complete a simple task, going from instruction to solution without lots of tedious middle steps.

Lyu wrote on X that the Rabbit sits in a sweet spot with “god-level design from Teenage Engineering, retro-culture resonating the good old fun tech era” and the best price on the market for a device of its type.

It isn’t designed to be a replacement for the smartphone, but rather a way to show an audience “it can do some things better than a phone and much faster”. It can also be trained to carry out other tasks not in its default programming through an intuitive web interface.

Since its launch a number of AI companies including artificial intelligence search engine Perplexity have expressed interest in building services for the device.

When will the Rabbit r1 be available?

The Rabbit r1 is available for pre-order in the U.S. now, with global sales expected to open at some point in the next few months. The first devices will ship around Easter.

The pocket-sized assistant is priced at $199 with no ongoing subscription fee, unlike its closest competitor the Humane AI Pin which starts at $699 and includes a monthly subscription.

So, is this product solving a problem that doesn’t really exist?  I put it in the same category as a self driving car- is it a solution looking for a problem? I have so much power in my pocket via cellphone; am I too lazy or stupid to figure out which application to assign a task that I need to only communicate via voice and expect (need) to have the device figure out how to grant my request? For sure, it is smaller, and fits in my pockets better.  It is inexpensive, but what am I giving up as far as access to apps, durability, longevity (at least we have been accustomed to two years), compatibility, and on ad infinitum. Phones keep growing, but there are enough choices on the market to solve anybody’s size needs.  IMHO, this device presents a scary concept to me, as I feel that I would lose control of choices and put my needs in the hands(?) of the AI’s programmer and training.  For $199, how much could there really be there?  Why would I really want only 4G support?  Note that the 2024 non-CES price is about $450.  I still have many reservations and what-ifs for my own acceptance. -GBL

A Second Opinion

Why the Rabbit R1 AI Is the Most Exciting Thing Since the iPhone

Finally, an alternative to the app-based phone

By Charlie Sorrel

Published on January 15, 2024, Lifewire

Trending Videos

  • The Rabbit R1 is an AI-based handheld that uses your apps for you.
  • It uses an AI “Large Action Model” trained on gazillions of app interfaces.
  • No ads, no distractions, no dark patterns. Just tell it what you want, and it does it.

The Rabbit R1 might look like a Tamagotchi from the TV series Severance, but it might also be the first credible alternative to the super-dominant iOS/Android paradigm for mobile devices.

The R1 is how you might imagine a pocket computer if you had never seen a smartphone. You just tell it what you want it to do, and it does it, formatting the results into a beautifully simple card on the unit’s small screen. And it can work with pretty much any app in existence because it uses AI to interact with them, like a human with infinite patience.

“The large action model promises a more intuitive, voice and AI-centric interaction, signaling a move away from the traditional screen-focused paradigm. This shift could potentially reduce screen time and make technology more accessible to a broader audience, including users with visual impairments or tech literacy challenges,” Brian Prince, founder and CEO of AI educational platform Top AI Tools, told Lifewire via email.

The Future We Were Promised

The R1, designed in partnership with Teenage Engineering, is pretty much entirely voice-based, and it doesn’t run apps. Instead, you push the side button to give it instructions, like talking into a walkie-talkie. Going by the demo video, it’s fast and eerily accurate in understanding your commands.

That’s because it uses AI, although not in the way we’ve become accustomed to. Instead of spitting out convincing garbage using a Large Language Model like other “AI” services, the R1 uses what Rabbit calls a Large Action Model, or LAM. This has been trained not on the words or pictures of the internet (like a large language model—LLM) but on the user interfaces of its web apps.

This means it can understand and use any website on your behalf, which is pretty nuts. Demo examples include booking an entire vacation, including flights and hotels, and creating an itinerary. You can tell it to order an Uber to take you home, order dinner, or play music. It will then interface with those web apps and websites for you and tell you about the result.

It sounds gimmicky, but you could think of it as fulfilling the promise of computers, the kind you see in Star Trek, the kind that just get stuff done. You never see Nyota Uhura or Wesley Crusher struggling with TripIt or juggling apps to hail the enemy vessel.

We’ve gotten used to websites that try to trick us into signing up for mailing lists while we buy, ads that track us across the web, and a newer, more annoying interface for every new service that we are forced to learn just to order a pizza. We’re also used to picking up the phone to, say, check the time of an appointment and then getting distracted for half an hour by Instagram or TikTok, having not even checked the calendar app.

This is the kind of AI we can get behind. Instead of ripping off the creations of zillions of humans and then using them to make up incorrect “facts,” Rabbit’s LAM is way more focused. It’s been trained to use websites and to parse instructions from your spoken commands. Those are both things that AI should be good at. And yes, it will answer questions with facts, but as far as I can see, it doesn’t use an LLM for that. It’s just a plain old web search.

Take Two Devices Into the World?

Many of the replies I have had to my request for comments on the Rabbit R1 focus on using two devices. “Who wants to use this and their phone,” they ask. “This should just be an app.” To which I will answer, equally subjectively, that I would totally take this out with me instead of my phone. I’d be able to do all of what I need to do when I’m out—check routes, listen to music or podcasts, and—presumably—send people messages, all without having to deal with apps. It’s

If you soak your brain in TikToks while riding the metro, then this is not for you. But as Rabbit’s CEO and founder Jesse Lyu says in his keynote, the R1 isn’t meant to replace your phone. Or not yet, anyway. What it is, though, is a first glimpse at an alternative to the touch-screen, app-based hegemony of Android and iOS. Today, all phones are essentially the same and work the same way. The R1 still has to prove that it can do what it says it can in real life, but it seems that some kind of AI-driven interface is going to be what finally usurps the app-based model.

“Samsung recently teased Galaxy AI, which is similar, but it won’t be able to perform the tasks that R1 Rabbit can. However, likely, smartphones will soon include AI,” AI software developer Rokas Jurkėnas told Lifewire via email.

And here’s the kicker. It’s only going to cost you $199 to find out how useful the Rabbit R1 can be—with no subscriptions required.

So maybe I was wrong.  I doubt it though.  I have a long-time engineering and software background, so I am not like most consumers- I like to retain control of my destiny and decisions.  Trust is not a forgone conclusion on my part. I will live with my one poor choices and frustrations rather than put my trust in another engineer or programmer. I’m the one that will never accept self driving cars because I would have to put my trust in someone else’s unproven skills.  For some people, they are ready to put their trust in any device that make’s their life easier.  I will not be returning to 4G communication.  I like 5G just fine. I like to read my news and weather on my phone, since real newspapers are quickly becoming scarce.  Will I carry multiple devices? Probably not. Should this become an option on the cellphone of my choice, I will most likely try out the capabilities, but I can also opt to turn it off. Is the Rabbit R1 a foot-in-the-door of progress? Probably.  Am I ready for it?  Probably not. Perhaps a better name would be Destiny. – GBL

General AI: 

https://arxiv.org/abs/2310.02255

New Business: Optics in AI and Quantum Computing:

https://www.lightreading.com/optical-networking/looking-ahead-to-2024-gen-ai-creates-optical-opportunities#close-modal

https://optics.org/news/14/11/15

Conclusion:

While any conclusions on the current state-of-affairs in artificial Intelligence would be speculative, what remains clear is the urgency of review and investigation into the ‘what if’ scenarios of the current artificial intelligence environment. My biggest question is when common interests in related technologies will become involved in developing the inherent advantages of the suggested neuromorphic optical architecture. As a ‘Math Guy’, I continue to rely on the standard ‘Definition-Theorem-Proof’ approach in these matters. An intuitive geometric proof can be seen in the optical interconnect work (and page below- with a major site update). Please visit the site and review these many advantages thoroughly. Avid readers of this newsletter will note that we have converted from a weekly newsletter to a monthly release! We will be supplementing this great volume including your curated recommendations. ‘Much more to come-and continue with new volumes thereafter. Thanks for reading and contributing and best regards!

https://worldview4d.com/

Leave a Comment

Your email address will not be published. Required fields are marked *

While reasonable effort has been expended on double-checking facts, your authors and everyone associated with this website make no claims as to accuracy. We also try to avoid trademark or copyright infringements. If you feel something should be removed, please email contact@worldview4d.com.