- Everyday AI
- Posts
- How AI Agents Can Bridge the Gap to the Future of Enterprise Work
How AI Agents Can Bridge the Gap to the Future of Enterprise Work
OpenAI gets copyright claims dismissed, Google unveils an AI learning assistant, Mistral challenges OpenAI with new API, Baidu releases AI smart glasses and more!
Everyday AI
November 08, 2024
š Subscribe Here | š£ Hire Us To Speak | š¤ Partner with Us | š¤ Grow with GenAI
Sponsored by
āMy advice for CEOs: figure out what problem youāre trying to solve, then weāll see how AI can support that,ā says Delphine Zurkiya, a senior partner at McKinsey & Company.
She joins the WorkLab podcast to discuss how leaders should approach deploying AI and reskilling teams for the new era of work.
Learn more on WorkLab, available wherever you get your podcasts.
Outsmart The Future
Today in Everyday AI
6 minute read
š Daily Podcast Episode: Why are AI agents booming right now? What are they? We explain how they can boost enterprise work. Give it a listen.
šµļøāāļø Fresh Finds: Microsoft Outlook gets AI themes, Sune teases V4, Amazon may invest more in Anthropic and Lionsgateās CEO speaks on Runway partnership. Read on for Fresh Finds.
š Byte Sized Daily AI News: OpenAI gets copyright claims dismissed, Google unveils an AI learning assistant, Mistral challenges OpenAI with a new API and Baidu releases AI smart glasses. For that and more, read on for Byte Sized News.
š AI In 5: Google AI image model Imagen 3 has flown under the radar. We dive in and see how it stacks up against the competition. TBH, weāre kinda digging this. See it here
š§ Learn & Leveraging AI: Are AI agents the missing piece you didnāt know your business needed? We break down how they can help. Keep reading for that!
ā©ļø Donāt miss out: Did you miss our last newsletter? We talked about Microsoft integrating Copilot Pro AI features into Microsoft 365, Donald Trumpās AI plans and Anthropic partnering with Palantir and AWS for Defense AI. Check it here!
How AI Agents Can Bridge the Gap to the Future of Enterprise Work š§
Why AI agents? And why now?
You've prolly been seeing all the buzz around AI agents lately.
Same.
Here's the thing, though. There's more to them than meets the eye.
Scott Beechuk joins us to dive in deep and tell ya what you need to know.
Join the conversation and ask Jordan and Scott questions on AI here.
Also on the pod today:
ā¢ Current State of AI Agents š¤
ā¢ Humansā Roles in AI Integrationš¶
ā¢ Challenges in AI to AI Interactions š
Itāll be worth your 33 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Hereās our favorite AI finds from across the web:
New AI Tool Spotlight ā Latta finds and solves bugs automatically, Invofox turns files into verified data and CapCut Commerce is end-to-end content creation for commerce.
Microsoft ā Microsoft Outlook now has dynamic AI-powered themes.
Amazon ā Amazon may increase its investment in Anthropic under one condition.
Google ā Google has launched its AI-powered video maker, Vids, across multiple Workspace editions.
NVIDIA - Jensen Huang to discuss AIās future with Masayoshi Son at AI Summit Japan.
AI in Media ā Lionsgateās CEO spoke on the recent AI partnership with Runway.
AI Music ā Suno teased the next iteration of its AI music generator, V4.
v4 is coming soon š§
ā Suno (@sunomusic)
6:55 PM ā¢ Nov 8, 2024
AI Models - Duolingoās new AI chatbot replaced its CEO on an investor call to show its capabilities.
Read This ā A cyber attack warning has been issued as hackers use AI and gmail.
AI Study ā IBM released its āAI in Actionā report, a goldmine for companies looking for key AI in business insights.
Agentic AI ā Salesforce is reportedly hiring 1,000 salespeople to push its agentic AI solution, Agentforce.
1. Court Dismisses Copyright Claims Against OpenAI š§āāļø
In a significant ruling, the Southern District of New York has dismissed a copyright lawsuit from Raw Story Media and AlterNet Media against OpenAI, stating the plaintiffs failed to prove any actual harm from the alleged use of their articles in training AI models. Judge Colleen McMahon highlighted the challenges in demonstrating concrete injuries in cases involving generative AI, where outputs are synthesized rather than copied verbatim.
This decision could shape future copyright claims as it emphasizes the need for clear evidence of harm, leaving content creators grappling with how to protect their work in an increasingly AI-driven world.
2. Google Unveils AI-Powered Learning Companion š§āš«ļø
Google has launched "Learn About," an AI-driven tool designed to transform how users engage with complex topics. This innovative platform allows users to ask questions, upload materials, or explore curated topics, providing tailored responses that cater to individual knowledge levels.
By creating interactive guides and resources, Learn About aims to support students and lifelong learners in mastering challenging concepts with ease.
3. Mistral AI Unveils New Moderation API for Safer Content š
French startup Mistral AI has launched a content moderation API that aims to compete with industry giants like OpenAI. This new tool, powered by their advanced Ministral 8B model, analyzes potentially harmful content across nine categories and supports 11 languages, positioning Mistral as a formidable player in the multilingual moderation space.
With recent collaborations with major companies like Microsoft Azure and SAP, Mistral is not just making waves but is also setting the stage for a safer AI landscape, particularly appealing to businesses navigating strict European data regulations.
4. Baidu's AI Glasses Set to Challenge Meta's Ray-Bans š¶
Baidu Inc. is gearing up to unveil its AI-powered smart glasses at the upcoming Baidu World event in Shanghai, marking a significant push into the hardware arena as it vies with Meta's successful Ray-Bans. The glasses, equipped with built-in cameras and voice interaction capabilities via Baidu's Ernie model, are expected to launch by early 2025 at a price point lower than Metaās $299 offering.
This launch comes amid heightened competition in the AI space, where Baidu is striving to reclaim its position against rivals like ByteDance, which has recently eclipsed its chatbot in popularity.
5. ByteDance's New AI Photos to Videos š¤Æ
ByteDance has just unveiled its "X-Portrait 2," a system that can transform still photos into incredibly lifelike video performances, capturing nuanced emotions that rival real actors. This innovation comes at a critical time as digital misinformation looms large, raising questions about authenticity in the age of AI-generated content.
With its vast TikTok database fueling this technology, ByteDance is poised to revolutionize industries like animation and video production, potentially replacing expensive motion capture setups with just a single photograph.
Is Google's Imagen 3 BETTER than MidJourney?
Google image generation model Imagen 3 hasnāt really been talked about.
Is it better than DALL-E 3 or Midjourney?
We show you how to access Imagen 3 and give a head-to-head comparison.
Check out today's AI in 5.
š¦¾How You Can Leverage:
You ever ponder about how certain repetitive human interactions are so insanely inefficient that we overlook just howā¦ insanely inefficient they are?
Like, how maybe some human jobs could be better off handled by AI agents and everyone would be happier and better off?
Case in point ā thereās countless call center jobs in the world where all the human agents do is help other humans reset their passwords.
Repetitive job for the human call center operator.
Annoyingly long wait times for the human who forgot their password.
Yuck.
Scott Beechuk sees a better way ā networks of AI agents.
Heās a Partner at Norwest Venture Partners, and his company invests in multiple companies using AI agents to cut down on mundane interactions and (hopefully) make way for more meaningful connections.
And nowās a gasoline-induced flaming hot time to be involved in anything AI agents.
In the past few weeks, weāve seen both announcements and releases in the AI agents space from Anthropic, Microsoft, Salesforce and others.
So why now?
And how can companies balance the immense potential of AI agents with the potential peril they bring?
Buckle up, humans.
Scott is taking us to school on how AI agents are gonna be impacting enterprise work.
Hereās what you need to know.
1 ā Convergence is happening nowš”
Wait, so why is all of this happening now?
After all, we heard some āAI agentā rumblings years ago, but nothing ever seemed to materialize.
Yet in the last few weeks, weāve seen robust AI agent offerings from some of the largest companies in the world.
So whatās changed?
First, companies like OpenAI are making it easier to use their technology.
Second, the learning curve and requirements for building AI agents is shockingly low and capable agentic systems are able to pop up over night.
Lastly, Scott said AI agents now have the ability to connect with countless third party enterprise softwares and systems, but can now actually execute tasks with those systems.
Try this:
Even weāve been buzzing a bit more about AI agents here at Everyday AI.
Why?
Companies like Salesforce and Microsoft bringing autonomous AI agents to market in 2024 legitimizes this tech that was once more scientific fiction than reality.
If you want a more well-rounded view on the AI agent space, make sure to check out this episode.
2 ā Agents coming for Tier 1 š
Scott has already seen this happen ā AI agents are coming for Tier 1 roles.
And itās not always in a āAI agent taking a human jobā type of way, but is oftentimes just an AI agent taking on pieces of responsibility that humans would normally take on.
So whatās a Tier 1 support role?
Think of something like an entry level customer service position or even a Sales Development Rep.
So many of the first-touch connections that companies have with their customers are routine, repetitive and donāt always require much finesse from a human.
(Like instructing someone how to reset their password.)
Try this:
Scott walked us through an example of this from Replicant, an AI agent company that Norwest invests in.
He gave the example that if you call AAA insurance in North America, thereās a good chance your first conversation might actually be with an AI agent that sounds and reacts more like a human than a tele-operator.
You can go read more about Replicantās partnership with AAA here, and then take a deep dive on how AI agents are already disrupting customer service and sales industries.
3 ā Future agent networks need rules ā
Itās a weird world weāre living in yāall.
Likeā¦ thereās actual AI agents in the wild right now doing roles traditionally reserved for humans.
AI agents of today can talk like a human, listen intently, respond in low latency (faster than the average human in some cases!), and do it all with access to your companyās information.
Yikes.
Right now, we have live, working individual AI agents. But, that will soon change as AI agent networks become mainstream.
Cool right?
Yeah!
Dangerous?
Maybe.
But Scott cautioned this is where business leaders have to balance the power of AI agents (or groups of them!) with the current realities of large language model-powered agents.
Right now, agents can talk to each other and solve problems and execute tasks. But if Agent 1 hallucinates a bit, and Agent 2 hallucinates even furtherā¦. You can only imagine the mess youāll be in when agents 903 and 904 are executing actions on your companyās behalf.
Double yikes.
Try this:
You remember that telephone game?
One kid whispers a funny secret into another kidās ear.
After the message gets passed to 9 or 10 kids, itās completely different than the original message.
Scott says thatās what could happen with AI agent networks.
He said companies who use agentic AI networks first need to develop a tracing system like developers use with APIs.
The issue is thereās no standard language for agentic AI networks. Normally, network admins and devs can go through logs of their tech stacks, tracing API calls to find bugs.
With LLM-powered AI agents, not so much.
If youāre still new to AI agent networks (and arenāt we all?!) ZDNet has a great breakdown of what it all means.
ā
Numbers to watch
$1 Million
The first artwork painted by a humanoid robot to sell at an auction sold for $1 million.
Now This ā¦
Let us know your thoughts!
Vote to see live results
Do you attend our livestreams?Every weekday, we bring you fresh AI insights, exclusive interviews, and breaking news with our Everyday AI livestream. |
Reply