www.trinetix.com Open in urlscan Pro
3.167.88.23  Public Scan

Form analysis
0 forms found in the DOM

Text Content

About
Company
Our Team
what we do
Enterprise Operations Enablement
Productized Business Services
Innovation & Tech Excellence
featured services
AI and Generative AI Services
Experience Design
Enterprise Software Development
Intelligent Automation
Data and Analytics
Blockchain Solutions Engineering
Digital Workplace
Intelligent Digital Assistants
Mobile App Development
Cloud Enablement
Industries
Professional Services
Logistics
Financial Services
Healthcare
Agriculture
Case Studies
Insights
Get in Touch

 1. Trinetix
 2. Insights
 3. LLM vs LAM or The Dangers of Marketing Hype Investors Should Know About
    


LLM VS LAM OR THE DANGERS OF MARKETING HYPE INVESTORS SHOULD KNOW ABOUT

AI/ML
7.11.24
Dmytro Ivanov
MACHINE LEARNING ENGINEER
Alina Ampilogova
COMMUNICATIONS MANAGER

Rapid digital transformation and the latest technological disruptions put
entrepreneurs and business leaders in front of new game-changers and potential
breakthroughs. As 45% of companies with a strong innovation culture reported
successfully transforming their processes without stalling and 63% documented
increased product and service quality, it becomes clear that exemplary business
models of tomorrow will be digital, high-tech, and AI-powered.


However, decision-makers are bound to encounter certain risks in their pursuit
of opportunities. In this article, we'll discuss these dangers using the recent
Large Action Models (LAM) case as an example. 

Table of contents
 * What is LAM: the importance of investigating disruptions
 * What to consider when making your choice?
   * Question everything
   * Search for the root of the hype
   * Explore existing technologies first
   * Wait for the dust to settle
 * Navigating the hype: how to find the right technology for your project

Table of contents
 * What is LAM: the importance of investigating disruptions
 * What to consider when making your choice?
   * Question everything
   * Search for the root of the hype
   * Explore existing technologies first
   * Wait for the dust to settle
 * Navigating the hype: how to find the right technology for your project


WHAT IS LAM: THE IMPORTANCE OF INVESTIGATING DISRUPTIONS

The actual answer to "What is LAM?" is relatively short — large action models
don’t exist. 

The term LAM entered the tech community field when the AI startup company Rabbit
started promoting their AI device for the B2C audience. According to Rabbit
campaigns, their product R1 runs on an innovative Rabbit OS powered by a new
artificial intelligence capable of making complex decisions and acting as an
intelligent assistant for everyday tasks. It allows Rabbit OS to help users with
activities such as ordering groceries, reserving tables at a restaurant, playing
music on request, and sending messages.

> Essentially, the first questions start right at the marketing promises. Why do
> I need a separate AI-powered device for playing music when we already have
> Alexa and Siri to do that? Next, the booking and food ordering options – does
> that mean that AI has access to the users' credit card and banking
> information? Such a feature requires immaculate safety and security protocols
> and detailed user instructions. However, the developers withhold any
> information on how they will ensure security, which should raise suspicions
> already.
> Dmytro Ivanov, Machine Learning Engineer at Trinetix

Patented technology and professional secrets exist, but so does responsibility
to end users. If a company markets a product that is claimed to interact with
personal information and perform actions on the user's behalf, it has to ensure
that such features won't be exploited with malicious intent. 

While AI facilitates risk management and enhances security, it can also be used
for less legal activities. If a technology or model isn't properly tested for
exploits, then these exploits will ultimately cause harm. 

> Let's say an AI model is really trained to access a restaurant website,
> reserve a table, and pay in advance. However, what are the precautions against
> a random person getting their hands on the device and commanding the AI to
> make multiple orders until the user's credit card is depleted? If the system
> is based on voice recognition, can it be tweaked with the help of the user's
> voice recording? In our new reality, it's important to consider the benefits
> driven by AI and the dangers and blindspots. Especially when you are offered a
> product that is supposed to be a part of your everyday life.
> Dmytro Ivanov, Machine Learning Engineer at Trinetix

In the context of these risks, it's remarkable that Rabbit, as the initial
promoter and creator of LAM, not only refused to address these concerns but also
avoided providing any explanation regarding how large action models work.

Why is it a red flag? 



 * Credible developers want to educate end users
   No matter how complicated the technology behind the product is, it's
   important to explain it to end users for safe interactions, maximum
   productivity, and seamless communication. Failure to provide important
   information on how the technology works can lead to financial or reputational
   damage, which will affect the developers' reputation.
   
   
   
 * Dishonest developers rely on marketing hype
   Words like "innovative," "disruptive," or "sensational" should never be
   enough for a decision-maker to make an investment. A product creator
   committed to long-term sales and support is expected to be prepared for a
   detailed explanation of features, comparing them to existing products and
   proving their superiority in some real use cases. However, when marketing
   campaigns are the only driving power behind the product name, it should be
   the reason for concern.
   
   
   
   
 * Avoiding direct questions is never a good sign
   The refusal to give clear instructions or explanations shows either a lack of
   confidence or intentional deception of the end users. In the latter case,
   dishonest promoters usually provide vague or self-contradictory insights,
   encouraging potential customers to buy and try the product for
   themselves—because their goal is to gain profit, not to provide long-term
   value.



In the case of R1, the overhyped marketing campaign has already been followed by
a disappointing reveal. Instead of independent decision-making AI, Rabbit OS
turned out to be a Large Language Model created by Open AI and using Playwright
web automation tools. 

> For that reason, the entire LLM vs LAM debate makes no sense right now. The
> existing Large Action Model is actually a Large Language Model agent under the
> guise of innovation. While there is nothing wrong with creating an intelligent
> assistant that leverages LLM agents for managing different activities and
> organizing routines, it is certainly a bad practice to promote such a product
> as an entirely new, revolutionary invention.
> Dmytro Ivanov, Machine Learning Engineer at Trinetix

This particular case serves as a reminder to stay aware and ready to fact-check
any disruption that enters public discussion. In the case of R1, buyers paid for
functionalities and capabilities they never received. However, such a type of
deceptive marketing isn’t going to limit itself to the B2C field only–and poor
investment decisions in the B2B area don’t punish just one group. The entire
enterprise has to deal with the aftermath.  

Given that in the wake of Rabbit promotion campaigns, several online media
started debating the application of large action models for businesses, it makes
sense to outline several steps executives and decision-makers should take before
investing in a technology.

Take a deep dive into the functions, use cases, and potential of large language
models
Read more


WHAT TO CONSIDER WHEN MAKING YOUR CHOICE?

There is one thing that can be said with certainty: the case of large action
models is the first of many. There will be many more loud entrances of products
claimed to be revolutionary and a must-have for businesses or consumers.
Sometimes, these statements would be backed with actual features and measurable
potential. However, to distinguish them from the innovation consisting purely of
marketing hype and promises, investors need to do the following.


QUESTION EVERYTHING

Asking the right questions can instantly remove any doubt regarding what kind of
technology to choose. For example, if investors seriously ponder the LLM vs LAM
comparison, they can settle it easily by diving into how each model works. They
will get a detailed explanation of how large action models work, what they can
and can't do, and how they can be used.

At the same time, they won't find any similar information regarding large action
models—because such models have never been applied anywhere, and thus, it's
impossible to prove their superiority or even their reality. Accordingly, seeing
the real pros of LLM and finding no feasible pros of LAM, investors will be
confident in their choice.



Such an approach can and should be applied to any novelty emerging on the
investors' radars. Being inquisitive allows decision-making groups to organize
priorities and dissect innovation in the context of real issues and challenges. 

HOW DOES IT WORK?

What are the principles behind the technology? If it’s an AI model, how is it
trained? How much resources and time does it take to train it? What kind of data
can it work with? How does it interact with other services?

ARE THERE ANY USE CASES?

Are there any companies using this innovation right now? Is there any positive
feedback? Has this technology been leveraged in my area of interest?

HOW IS IT DIFFERENT FROM EXISTING OFFERINGS?

How is this technology better than existing options? How does it perform in
comparison to the alternatives? Are there any benefits to keep in mind?

HOW DOES IT ENSURE DATA SECURITY?

Can the technology be exploited? What are the security protocols for user and
business data? Are there any regulations to know about?

HOW CAN IT SOLVE MY SPECIFIC BUSINESS NEEDS?

Will this innovation be useful for my enterprise? What kind of advantage can be
expected? What is the estimated ROI? How will it affect the operating expense
ratio?

> While these questions may seem simple and obvious, they are sometimes easily
> dwarfed by a natural wish to find the miracle solution or an instant
> competitive edge. It is normal to want to find something faster, better, and
> stronger than whatever your competitors have. But at the same time, nobody
> wants to find themselves a beta tester of a product that has yet to prove its
> value. So, looking for evidence, testimonials, and details should be an
> imperative.
> Dmytro Ivanov, Machine Learning Engineer at Trinetix


SEARCH FOR THE ROOT OF THE HYPE

It’s essential to identify who initially generated the latest trend and how it
was done. Was it a well-known tech company with multiple successful projects
behind its back? Was it a trusted industry voice sharing their honest opinion?
Or was it a promotional company for a product that hasn’t even been used yet? 

While many ambitious and innovative startups are coming up with new perspectives
on everyday approaches to technology, there are also projects that one Apple or
Google update from losing relevance. Startups that want to make a real
difference for users and businesses are aware of this—so they refrain from
setting unrealistic expectations and focus on few yet impactful benefits.
Additionally, they aim to partner with companies that have an influence in the
industry, gaining credentials and testimonials that hold weight.

Exploring AI in test automation: how to focus on value and leave the hype behind
Read more


EXPLORE EXISTING TECHNOLOGIES FIRST

Rapid tech evolution and the emergence of disruptions can be overwhelming and
create a false impression that a silver bullet is somewhere out there, at the
very threshold of emerging tech trends. 

However, searching for or waiting for such a bullet usually leads to either
missed opportunities or dissatisfying results when it turns out that the adopted
innovation has no synergy with enterprise operations. For that reason, it's
important for executives to stay focused on the technologies that have already
proven themselves, have multiple use cases, and are part of modern business
models.

Technology's availability and relative mundanity doesn't automatically imply it
has nothing revolutionary to offer. It all depends on the specific goals and
issues an enterprise is dealing with.

> There is one simple truth—no technology is better than others. The diversity
> of innovation enables enterprises to build ecosystems tailored to their
> business vision, company culture, and needs. This is the most important aspect
> of the modern digital age.
> Dmytro Ivanov, Machine Learning Engineer at Trinetix


WAIT FOR THE DUST TO SETTLE

Every disruption, real or not, always enters a market with a bang, putting its
advantages and strong points forward. It takes a while for complications,
issues, and negative feedback to emerge. Despite its powerful influence on the
way companies operate and services work, artificial intelligence is not an
exception to this rule. As AI platforms brought by such giants as Google and
Microsoft were reported to generate wrong recommendations and spread dangerous
misinformation, it becomes obvious that every novelty should be scrutinized and
given time to take root in reality, whether it came from a big-name brand or an
ambitious startup. 

This fact includes not just hype-based products but also tangible ones that,
despite their best efforts, have underwhelming features and low sales. 

The latter case doesn't mean there won't be any improvement—but it's better to
monitor the technology and its growth path safely without being caught in sudden
pitfalls and capacity gaps. Therefore, executives, particularly technology
executives, should be both proactive and patient, actively exploring new
opportunities brought by innovation while thoroughly investigating every trend
and its value.

Digital transformation: 9 signs your enterprise needs an upgrade
Read more


NAVIGATING THE HYPE: HOW TO FIND THE RIGHT TECHNOLOGY FOR YOUR PROJECT

There is no way to predict technology: solutions that were believed to be
impossible today can become part of everyday routine in less than 7 years. So,
the emergence of new AI models, more responsive AI systems, and many other truly
sensational disruptions is more than possible. 

However, despite the optimistic picture and the promise of opportunities, the
key to identifying the right fit for enterprise enablement is relying not on
trends but on the enterprise's specific needs and performance constraints. 

Another important factor in the choice of technology is expenses and profits.
Investors should have a clear and transparent vision of how much they are
willing to spend and how much they expect to gain.

> When a client comes to us with an intention to adopt a certain innovation, we
> start our cooperation with a discovery session, within which we ask a simple
> question “Do you really need this specific technology?”
> 
> In our work, our priority is helping our clients digitize their operations in
> a way that delivers maximum value and long-term outcomes. Because of that,
> sometimes it's not about what they think they need but about what they
> actually need. If there is an opportunity to adopt another technology that is
> easier to onboard and guaranteed to deliver similar value with lower expenses,
> we always suggest exploring it.
> Dmytro Ivanov, Machine Learning Engineer at Trinetix

Driving a competitive edge for our Fortune 500 partner through a strategic
AI-powered revamp
See case study

If you want to reinforce your plans with a 360° understanding of technology
trends and the opportunities you can glean, let’s chat. 

Within a strategic session with our experienced tech talents, you will receive
detailed recommendations regarding your individual business needs and growth
potential—and within a full technology partnership, you’ll have this potential
realized with maximum impact for your enterprise and industry.

Share:


RELATED INSIGHTS


THE WHAT, WHY, AND HOW OF LARGE LANGUAGE MODELS

AI/ML
9.18.23


AІ IN THE WORKPLACE: MANAGING ENTERPRISE DATA WITH CHATGPT

AI/ML
7.11.23


WHAT ARE LARGE ACTION MODELS AND HOW DO THEY WORK?

AI/ML
3.29.24


WHAT IS NATURAL LANGUAGE PROCESSING AND HOW DOES IT WORK?

Digital Assistants
3.1.24

Ready to explore
 tomorrow's potential?

Let’s get started

connect
Our Nashville office 1000 Health Park Drive, Brentwood, Tennessee

hello@trinetix.com

services
 * AI and Generative AI Services
 * Enterprise Software Development
 * Experience Design
 * Data and Analytics
 * Intelligent Automation
 * Digital Workplace
 * Blockchain Solutions Engineering
 * Mobile App Development
 * Intelligent Digital Assistants
 * Cloud Enablement
   

industries
 * Professional Services
 * Financial Services
 * Logistics
 * Healthcare
 * Agriculture
   

company
 * About Us
 * Services
 * Case Studies
 * Insights
 * Contact Us
 * CSR Report
 * Corporate Policies
   

Our brands
 * 
 * 
 * 
   


© 2024 Trinetix. All Rights Reserved.
Privacy NoticeTerms of UseCookie Policy