"I tested 400 AI Agents, these are the best" - Adam Silverman
Summary
TLDRIn this insightful podcast, Adam Silverman, founder of Agency, discusses the transformative impact of AI agents on various industries. With over 400 agents tested, Silverman highlights their efficiency in sales, research, and data enrichment. He emphasizes the potential of tools like Claude, Lutra AI, and Omni, and shares his vision for a future where AI seamlessly integrates with human tasks, enhancing productivity. The conversation also touches on the importance of staying updated with AI advancements for personal and professional growth, and the opportunities presented by fine-tuning AI models for specific tasks.
Takeaways
- π§βπΌ The speaker emphasizes the utility of AI agents in performing tasks overnight and while the user sleeps, highlighting the potential for increased productivity and efficiency in business operations.
- π The testing of over 400 AI agents has been conducted, with at least 50 actively working for the speaker's company, showcasing the extensive exploration of AI capabilities in various tasks.
- π The announcement of a $2.6 million fundraising under the name 'Agencies' indicates a significant investment in the development and integration of AI agents in business processes.
- π οΈ The speaker discusses the importance of integrating AI agents with various platforms and the potential for building complex, custom solutions for Fortune 500 companies.
- πΌ The concept of 'agent teams' on different platforms for various tasks is introduced, suggesting a strategy for leveraging specialized AI tools across different business functions.
- π‘ The introduction of 'Mindy', an AI agent that provides detailed briefings on meeting attendees, illustrates the practical applications of AI in enhancing preparation and efficiency in professional interactions.
- π The speaker predicts a future where AI agents become proactive, taking initiative in tasks without needing human direction, indicating a significant shift in AI capabilities.
- π Concerns about data privacy and security with AI agents are acknowledged, with a discussion on the potential for local AI models to address these issues.
- π The potential for open-source AI models to surpass closed models in performance and cost-effectiveness is considered, suggesting a shift in the AI industry landscape.
- π The speaker recommends actively learning about AI and staying updated with the latest developments, positioning oneself as an 'AI expert' within one's social or professional circle to unlock opportunities.
- π The importance of identifying and automating repetitive and mundane tasks with AI is stressed, advocating for a proactive approach to embracing AI to enhance personal and professional productivity.
Q & A
What is the significance of testing over 400 AI agents as mentioned in the script?
-The significance lies in the extensive evaluation of AI capabilities across various tasks, ensuring that the best agents are selected for different use cases, ultimately improving efficiency and productivity.
How does the speaker describe their experience with AI agents working overnight?
-The speaker describes it as a highly efficient experience where they can go to bed with agents working and wake up to find the agents have already been working all night, demonstrating the 24/7 operational capability of AI agents.
What is the role of 'Omni' in the context of AI sales agents mentioned in the script?
-'Omni' is an AI sales agent tool that excels in data enrichment and prospecting, helping to identify and gather information on potential sales leads effectively.
Can you explain the 'God Mode' tool and its capabilities as highlighted in the script?
-'God Mode' is an AI tool that the speaker finds incredibly useful. It can perform tasks such as finding every Fortune 500 CTO, enriching their data, and even building out full email sequences, all within its app.
What is the purpose of the 'Claude' tool that the speaker is obsessed with?
-The 'Claude' tool is used for a variety of workflows and is favored by the speaker for its versatility and effectiveness in handling multiple tasks, showcasing its utility in day-to-day operations.
What does the speaker announce regarding the funding and the future of their company?
-The speaker announces a successful fundraising round of over $2.6 million under the name 'agencies' (A.C). The funding will be used to help companies get their AI agents into production, indicating a pivot towards providing pre-built agents and custom solutions as needed.
What is the role of 'Mindy' in the speaker's daily routine?
-'Mindy' is an AI agent that analyzes the speaker's calendar and provides detailed reports on everyone they are meeting with, offering valuable data points and insights to prepare for meetings effectively.
How does the speaker view the future of AI in terms of handling repetitive tasks?
-The speaker believes AI is best suited for repetitive and less creative tasks, freeing up humans to focus on more enjoyable and creative work, thus enhancing overall productivity and job satisfaction.
What is the potential impact of AI agents on human jobs according to the speaker?
-The speaker suggests that AI agents will not replace human jobs but rather transform them. By handling the more mundane and repetitive aspects of work, AI agents will allow humans to focus on tasks that require creativity and reasoning, which are inherently human skills.
What is the speaker's view on the importance of being an 'AI guy' or leader in one's company?
-The speaker believes that becoming an AI leader or expert within a company is a strategic move that can future-proof one's job. It positions the individual as a valuable asset, especially in times of change or uncertainty, and can lead to promotions and increased job security.
How does the speaker suggest one should stay updated with the latest in AI?
-The speaker recommends following top people in the AI field on Twitter, subscribing to AI tool updates, and joining communities to stay informed. They also suggest creating a separate account for AI-related content to avoid information overload and maintain focus.
Outlines
π€ AI Agents in Business Automation
The speaker discusses the extensive testing of over 400 AI agents, emphasizing the efficiency gains from having multiple agents work autonomously around the clock. They highlight the potential of AI in various sectors, particularly in sales and research tools, mentioning specific agents like Omni for data enrichment and God mode for comprehensive data gathering and email sequences. The conversation includes the founder of Agency, Adam Silverman, who shares insights on AI agent use cases and announces a significant funding round for his company, indicating a strong belief in the future of AI agents in business processes.
π Advancing AI Assistants and Productivity Tools
The script delves into the use of AI assistants to enhance productivity, with a focus on tools like notaker apps that facilitate meaningful engagement during meetings by handling note-taking. The speaker expresses anticipation for a future where AI assistants have a comprehensive understanding of users' goals and work habits, offering real-time assistance and accountability. They also touch on the importance of privacy and data security, discussing the trade-offs between local AI models and cloud-based services, and the potential for AI to transform how individuals and companies operate.
π Open Source AI and the Future of Automation
This paragraph explores the rapid advancement of open-source AI models, which have begun to rival or even surpass closed models in terms of performance. The speaker speculates on the possibility of an open-source model becoming the best in the market, driven by companies like Meta. They also predict a future where AI models become so advanced that incremental improvements plateau, leading to a focus on fine-tuning and smaller, more specialized models. The potential for AI to become a standard tool for businesses, requiring specialized fine-tuning for specific tasks, is highlighted as a significant opportunity.
π Becoming an AI Leader in Your Industry
The speaker encourages individuals to become AI leaders within their respective fields by staying updated on the latest AI developments and integrating AI tools into their workflows. They share practical tips for learning about AI, such as following key influencers on Twitter, subscribing to AI tool updates, and creating dedicated email accounts for AI-related communications. The importance of being proactive in adopting AI to secure one's position in a company and to stay relevant in the job market is underscored.
π Identifying Opportunities for AI Automation
The paragraph discusses the importance of identifying tasks that can be automated using AI, starting with the creation of a list of potential automation opportunities. The speaker advocates for a proactive approach to finding and implementing AI solutions, particularly in areas like financial and equity research, where AI can outperform human capabilities. They argue that embracing AI will enable individuals to focus on the most human aspects of their roles, such as creativity and reasoning, while AI handles the repetitive tasks.
π The Integration of Humans and AI in Task Completion
The final paragraph examines the limitations of current AI agents and the innovative approaches being taken to integrate human input when agents encounter tasks they cannot complete. Examples include agents that can pay humans to perform tasks or verify human actions, highlighting a future where AI and humans collaborate effectively. The speaker envisions a future where individuals manage large teams of AI agents, overseeing operations and leveraging AI to its fullest potential.
Mindmap
Keywords
π‘AI agents
π‘Data enrichment
π‘Sales tools
π‘Research tools
π‘Automation
π‘Productivity
π‘Fine-tuning
π‘Open-source models
π‘Proactive agents
π‘Local models
π‘HIPPA compliance
Highlights
Over 400 AI agents have been tested for various tasks, with at least 50 actively working to enhance productivity.
The founder of Agency, Adam Silverman, discusses the practical use of AI agents in businesses and their potential for revenue generation.
Sales and research are identified as common verticals where AI agents are performing exceptionally well.
Omni and God Mode are highlighted as powerful sales tools for data enrichment and prospecting.
Claude and Lutra AI are mentioned as favorites for workflow automation and their effectiveness in handling specific tasks.
A massive fundraise of over $2.6 million is announced for AGOS, formerly known as agencies, to support agent deployment in companies.
Integrations with key players like Crei, Autogen, and others are teased, indicating a move towards more complex, pre-built agents for enterprise use.
The importance of custom solutions when off-the-shelf agents do not meet specific business needs is emphasized.
Mindy, an AI tool, provides comprehensive briefings for meetings by analyzing calendars and offering detailed reports on meeting participants.
The potential of AI in taking over repetitive tasks, allowing humans to focus on more creative and engaging work, is discussed.
AI note-taking apps like Fathom are praised for their ability to make meetings more meaningful and reduce the need for manual note-taking.
The anticipation of a complete AI assistant that understands goals, weaknesses, and provides real-time reminders is shared.
Concerns about data privacy and the desire for local AI models due to security questions are addressed.
The role of AI in the healthcare industry and the importance of HIPAA compliance for AI tools are mentioned.
Open-source AI models like Lama 3.1 and 405 billion parameters models are gaining popularity and are on par with top closed models.
The future of AI is predicted to involve smaller, finely-tuned models that can perform niche tasks more reliably and at a lower cost.
Becoming an 'AI guy' or leader in your company or social circle can provide significant opportunities and job security.
Practical advice for learning about AI includes following top experts on Twitter, subscribing to AI tool newsletters, and joining communities.
The transcript concludes with an invitation to join the New Society for a personalized plan to start an AI business, even for beginners.
Transcripts
we've been in the weeds testing over 400
agents I go to bed and I have agents
doing work for me I wake up in the
morning my agents have been already
working all night we have probably at
least 50 agents working for us I go into
all these agents testing them with very
similar problems that I I think there's
450 agents in total that we've tested so
far there's some actually cool ones off
the shelf that I think people can use
this is a podcast with Adam Silverman
the founder of agency we talk about
which AI agents everyone should use how
companies are making millions with
agents and how you can go from beginner
to expert all right so Adam what are
some of the most common use cases for AI
agents you see I think there are some
agents that are working really well
across like every vertical so sales for
example there's some really incredible
sales tools some really incredible
research tools out there can you talk
about the sales agents more because the
closer it is to generating money the
more people are interested there are so
many sales agents out there there's Omni
aomni it's incredible for doing like
data enrichment for prospecting then we
have another one called God mode it is
so incredible I can literally go in have
have you tried God mode no but I have to
okay you have to I'll get you on the
beta but really awesome tool I I was
actually using yesterday I can go in and
say find me every Fortune 500 CTO find
me their LinkedIn their email do full
data enrichment on it and you can build
Out full email sequences all within
their app what is like your new one of
the new tools that's your favorite I'm
still like obsessed with Claude I've
been using it every day I'm still very
bullish on lutra AI so I'm using that
for a lot of workflows I go into all
these agents testing them with very
similar problems that I have and then I
find certain agents are really really
incredible at certain tasks I have these
kind of like little agent teams on every
platform um that I used for different
things speaking of Agents you have
massive announcement to make right yeah
so uh some awesome news work sites
announced our fund raise of over $2.6
million under the name agencies so
A.C we've been in the weeds testing over
400 agents and we're really really
excited to announce this fundraise to
really help companies make sure they can
get their agents into production so if I
understand correctly agob is going to be
you know one of the tools and agency is
going to be more like holding company
yeah it's it's really exciting so we
have like a bunch of announcements
coming up so we still have Integrations
directly with crei with autogen um we
still integrate with laying chain llama
index mulon all these uh players
directly through aops but now we also
have the ability to go out and build
really really complex agents that
basically any Fortune 500 company can
start using Day Zero oh so you want to
like pivot more towards like pre-built
agents like there are going to be some
agents off the shelf that we offer
companies are coming to us with the
craziest agent IDE as possible like
truly like full n10 workflows that even
a human team couldn't physically do but
if custom Solutions are required we're
very happy to to figure out how to
basically leverage our our expertise our
knowledge our team of incredible
engineers and go and build out these
amazing solutions for them how big is
your team right now so uh it depends if
you count all the agents good answer
good answer yeah I know because honestly
I think we have probably I'm thinking
like at least 50 agents working for us
and uh like I go to bed and I have
agents doing work for me I wake up in
the morning my agents have been already
working all night there's some actually
cool ones off the shelf that I think
people can use most of you are waiting
you're waiting for the right opportunity
the right business idea the right time
but if you keep on waiting you're going
to miss the a Evolution me and my team
have spent hundreds of hours testing
different ways to make money with AI and
I believe we found the best one in the
new Society we help you go from zero to
$1,000 a month with AI I even give you
the AI agent I've built so you can make
money with it all you have to do is
follow our system for getting client
that's it so if you have good work ethic
and want to make money with AI make sure
to join it's the first link in the
description there's one called Mindy
what it does is it looks at your
calendar and then it gives you a full
like almost report on everybody you're
meeting with and before I go into a
meeting it gives me basically every data
point I could ever imagine about a
person this is really helpful imagine if
it took me let's just say 30 minutes or
my assistant 30 minutes to go research
everybody meeting within a day and i'
meetings in a day yeah that would be
that would be four hours of time and now
this is something that before I even
wake up in the morning the first message
I look at in the morning is here all the
meetings you have here's you know quick
fact on everybody here's their LinkedIn
here's maybe the most recent news
article they're mentioned in here's some
data on them that you know could make
the conversation exciting and this is
just like this is off the shelf like
this is not like crazy stuff by any
means and the fact that this is only
going to get better with better llms and
like you know it's just like it's crazy
for people who are running lots of
meetings like this is and again it's
it's not like um you know this is this
is the boring part of the work right
like researching stuff like you know
looking people always think that um AI
agents will take jobs but it's more like
you know the the repetitive boring task
so you can do more of the fun stuff more
of the creative stuff oh for sure like
it sounds kind of crazy but I think the
president gets these types of briefings
and I almost have that every day and it
really it's like it's fundamentally
leveling up how I am I think like
another you know very easy example
anybody can start using today is like
one of these notaker apps that you know
exist in your calls so like you use like
fathom and there's a whole bunch of
different ones out there and it makes my
followup so much more meaningful it
makes me you know really engaged in the
meeting because I'm not worrying about
taking notes and I think this is the
type of thing where it's so easy to set
up like as long as your company allows
you to use them and long as long as like
you have permission from people to
record it's such a massive win I cannot
wait for the day when we have like a
complete agent assistant that just you
know has access to your computer knows
your goals knows what you should be
working on knows like your weaknesses
like where you tend to waste time right
and just like holds you accountable
gives you real time reminders maybe
we've like connected to like Apple watch
or some other variable or maybe like you
know AI glasses I think this is coming
relatively soon and I think this is
going to unlock so much like human
potential right people always like don't
like the word productivity right because
you know people outside of business they
just feel like it's a corporate term but
just like human potential just imagine
like how much you know stuff goes to
waste purely because people waste time
you know everybody can write a plan in
the morning but sticking to it is much
harder right so so I think these AI
tools will just make that much more
doable and I mean yeah as somebody who
likes to be productive instead of
wasting time I just cannot wait for that
100% like the folks that induced are
doing this I I've seen like some
incredible work on their side um there's
a few other companies I'm really bullish
on the work of open interpreter so
there's a guy Mike bird from there he's
doing amazing work and kilan they're all
doing amazing amazing work there um and
truly like I made a tweet about this a
while ago but what I want is basically
everything I ever do in my life to be
fully recorded and I want to basically
almost have like a coach saying hey Adam
don't worry I'll take this off your
plate you don't need to send that email
or that newsletter you were going to
send out I you know let me get the
structure for you you can just review it
and I think the the big aspect we're
going to see with agents and I think
like you know maybe going a bit a year
or two down the road and this is what
I'm bullish on is that you're gonna go
from being reactive where it's like a
human asking to do something to
proactive and that's going to be the
biggest shift possible so I think right
now everyone's battling and like one of
the reasons like we're really like
seeing a lot of success with AGS is
because everyone's battling with
reliability so right now like the fact
is even on the pro or sorry on the
reactive front um it's it's still very
important to make sure that this agent
can do the task correctly but do you
think it has to be like local or
something like encrypted because you
know if it's recording everything
there's like big security questions
there yeah I think the folks at
Limitless AI have been working on this
quite a bit um if they're listening to
this like we we're always happy to chat
I am very bullish on what their team's
working on um but I think that there
will be you know maybe it's an Azure
offering or maybe a gcp offering or an
AWS offering where okay we now know that
everything is being recorded it's you
know by the user at big companies
everything's already being recorded if
you work at one of these big
organizations you know every keystroke
you make every you know uh every website
you visit everything's already being
recorded so it's not that crazy I think
whether or not the organization wants to
give the power back to the user to
leverage that that's another question
and also just like in the physical world
like if you go into any big city whether
it's SF new York Dubai there's cameras
everywhere right so I think the like
privacy is becoming more of an illusion
maybe like if you move into the forest
or whatever but uh I think the what what
people really have trouble with is
having the AI companies like access to
the data right I mean obviously if you
work for some company it's kind of
understood that the company you know you
probably have a work laptop and you
cannot access certain websites but if
you're an entrepreneur or like you know
in a spare time people a lot of people
don't want open AI to have their data
right or anthropic what ever they they
are really like afraid of um not
necessarily like somebody knowing what
they're doing but you know be having
that trained on in the future and they
just want to like that Assurance I mean
that's why the interest in local agent
is so big even though like most of the
times it's just like the difference is
so like the locally run stuff is just so
bad and especially most people don't
have like the topof thee line MacBooks
so they can only run run like 8 billion
models and you know the difference
between 8 billion models or like CLA son
or gb40 is just massive right so it
doesn't even make sense logically to
want that but still the demand is huge
because of you know people want that
feeling that you know open AI or
whatever isn't taking their data for
sure and like we so aov just became
hypoc compliant and so now we're talking
to like a lot of these Healthcare
companies and what we're realizing is
there are certain industries that I
think will be ryant on local model it's
a lot easier to get a CTO and a CIO to
buy into let's run this locally then
even if every vendor along the entire
spectrum is hippoc compliant and
everything like that it could still be
like you know they don't want these
infractions to say the least they do not
want to have any issues they don't want
to like you know have any risk but I I
still think there's actually opportunity
for like the most Cutting Edge companies
in the world to leverage the best models
and still probably remain competitive in
the long run so we'll kind of see how
this plays over the next six months and
we're going to do another episode I I
know eventually and talk about okay like
local models have now far surpassed like
you know all all like uh you know close
Source models but we we'll see
ultimately what ends up happening oh you
also predicted one my questions on my
list so you know we've seen a massive R
of Open Source especially in the last
like two to three months um AI images
have caught up and maybe even surpassed
closed models right um open source has
gotten a massive win with Lama 3.1 405
billion is on par with the top closed
models like do you think oh there is a
going to be a day where the best model
is completely open source I think that's
the like dream of meta and I think
mistol and a few of these other
companies are really trying to to make
that happen I think that it's possible I
think that we're also going to see a
Nexus point where it's like a race down
to the bottom in terms of cost so I
think like we're seeing that right now I
think that there's going to be a certain
level where these models are so good
that we don't need really much more
incremental um improvements so we'll see
over the coming months I think that
you'll have a world where like local
models are good enough um I don't know
if they're going to be the best quite
yet um but one thing I am extremely
bullish on and we might have t chatt
about it previously on or offline can't
remember but I'm very bullish on small
models like extremely I think like the
future is in very small like action
models even when you think about an
agent you don't need you know 400
billion parameters to go and conduct
some research on the web and scrape some
data um I think one of the big things
that's going to happen is you're going
to start seeing a lot of fine tuning and
that's one of the things we've been
working on AG Ops is like fine-tuning on
the outputs building much smaller models
or or fine tuning on the outputs of
basically Cutting Edge models and
fine-tuning open source models this is I
think going to be a big paradigm shift
because you're going to have cheaper
cost you might be able to do it locally
um there's going to be I think a lot of
work done in the space over the next
year because right now we're just saying
you know they're large language models
let's give them as many parameters as
possible let's try and like basically
you know put every like you know
increased context window as much as
possible um I think this is great on
very generalized use cases but from
everything we're seeing with like large
corporations and Enterprises the reality
that's you know I think what's going to
be most beneficial is actually could we
just get very niche done extremely
reliably we don't need to you know know
all about um ancient Rome you know to
you know conduct analysis on financial
documents yeah and I think U you know
for those who are watching who like kind
of thinking of
opportunities essentially every company
is going to need that right whether it
is uh language model fine-tuned or image
model fine-tuned they're going to need
something specifically made for them and
most companies don't know how to do it
so you know if you learn the skill and
if you have a bit of you know
inventiveness entrepreneurship this is a
pretty good opportunity in the next I
think year or so for sure like I think
you need I I love time people this but
like be the one at your company who
brings up this technology be the one who
people are like asking questions about I
have one guy he messages me it's
hilarious I do these threads like almost
every week about like what's the latest
and greatest in AI agents and I post
them on Twitter and or and Linkedin as
well and he takes these threads and then
goes and makes a PowerPoint deck and
presents it to his management team every
week now saying hey here's the cutting
and it's so funny to me like literally
like he's just he's taken my work I'm so
happy for it it's the best thing ever
I'm so happy he reached out by the way
and uh he's he's presenting this to his
team now and it's one of these aspects I
look at I'm like this is crazy like he's
now becoming like an AI leader at his
company by literally going onto Twitter
and looking at my thread making a small
deck about it and now he is you know by
far like the AI like agent guy in his
company that you know the Seau is coming
to him and asking him questions so it's
that simple like it doesn't be like it's
not rocket science by any means it's
it's a complete hack I mean just from my
experience right the fact that I'm in AI
it like I'm getting access to people who
are far ahead of me in business and in
life because they're interested in AI
right everybody wants to know at least a
bit how it's going to impact their
business you know how they can leverage
it and being in the right field the
field I mean I don't think there's a
betteri better field better industry
than AI obviously but you know just
being in AI brings so many opportunities
so I think everybody watching while you
you might you know don't want to have
Ambitions to be like a public figure you
should still become an AI guy in your
social CLE because that's going to
unlock so many doors that you can't even
predict if you're watching this video
you're already doing like the right step
that's the first thing it's like you're
already like light years ahead than a
lot of people I know and I think that
the reality is it's just like it's not
hard like it's just like listen to a few
you know podcasts listen to some videos
form your own opinion I think like
understand what tools are out there test
the tools like to me if I was and this
might be a hot take but it's okay it's
like if I was thinking about going and
doing like a masters in engineering or
if I was going to go do like a uh you
know some type of it related role I
would just go on YouTube for three
months lock myself in a room watch every
video on like AI possible and I I'm
almost for certain you will come out
here light years ahead than going and
doing like an IT certificate or a master
like master and Engineering might be a
bit different but I I'm saying like this
is such a massive WID like it's just
like there's so much content out there
there's so much knowledge out there like
I give away like 99% of everything I
know for free and like I'm just like I
see something cool I'll repost it I'll
share it I I'm happy to you know kind of
consolidate it so it's easy for people
to understand but the biggest factor is
like there's so much out there and even
like you know going and joining School
communities and finding like people who
are intelligent and are thinking about
in the same way or even like creating
group chats with you know on the cost
side it's not very high to to make these
like kind of think tanks uh for yourself
oh yeah for sure and I mean now is the
time really you know in the future when
like your Uber driver is talking about
it and everybody knows about it it's
going to be too late like now everybody
watching this really has the opportunity
to be kind of that AI Ford leader you
know whether it's in the region in their
company doesn't matter to what scale but
you know why not for anybody who's like
looking to learn more about like Ai and
AI agents and everything like that um if
you go to A.C our new website so agency
um you will go and click on like there's
a tab called agents you will find a list
of like 400 I think there's 450 agents
in total that we've tested so far and
it's totally free um I highly recommend
checking it out it's like the easiest
place maybe we can link it in maybe
we'll Link in the show notes as well but
this is like it's such lwh hanging fruit
for people it's like go explore it like
instead of like going uh you know one
day if it's raining outside or something
like that just like go like take that
list and just start clicking on a bunch
of the links and start looking at you
know what's working what's not what you
think is cool another thing I love to do
is go make an email account that is just
for like AI tools and go and subscribe
for all these players and what you'll
get is you like anytime you want a good
update on everything that's going on
just go and scroll through that inbox I
don't think you have to be an expert you
don't need to be an engineer you don't
need to have spent like 10 years in it
uh all you have to do is just kind of
lean in a bit test the stuff out and
kind of see see what works for you man
such a good advice this I love like
these practical tips you know that just
like have instant effect how are you
yourself learning about AI like we love
Twitter like I honestly my co-founder
Alex and I like in we we really just see
it as like kind of for whatever reason
in the AI space it's where we're seeing
the latest and greatest news
um I also do a lot of social listening
so I have agents that are out there on
the internet collecting data and then
giving me like report every day on like
what's going on what companies are
getting funded um and if anybody's
interested like I maybe I'll be able to
like kind of share this agent with you
yeah that would be a good video actually
yeah we can we can do we'll chat about
that offline but uh okay I think from my
perspective it's really really
interesting that leverage like the tools
to figure out the latest and greatest
but also you got to just be in the right
circles so following certain people will
basically guarantee you're getting the
latest and greatest news but at the end
of the day there's certain people who
are like you know like yourself um we're
trying to do a lot of like kind of this
thought leadership as well and it's just
a matter of kind of seeing everything
making it really easy for people to
understand if you're not like a big
Twitter user you don't need to like post
I think that's one thing a lot of people
are scared of doing like you don't need
to actively be posting all you have to
do is just create your own you know uh
account with whatever name on it follow
a couple of the top people and from
there your feed will start getting
populated with all the right content
maybe it is the move to create a
separate account because because it is
so easy to waste time on Twitter because
curate curating your feed is so hard
like constantly muting people like
muting words it it should have shouldn't
be that hard but you know everybody who
kind of is more active deals with this
problem of like I want interesting AI
stuff only right I mean I would pay for
that if I could just do that manually
instead of like trying to having to
update my muted word list constantly but
still like most of the tweets I see are
not like productive interesting AI stuff
I Me Maybe you can give me some tips
because you know so there's you can
create lists which is not a bad thing um
but from my perspective I think the best
thing is just following the right people
and then like looking through the
following like the For You tab or I
don't know exactly what it's called on
Twitter but like that is great I think
that's it's like really great um to find
content um but I think like like
following the right people is the most
critical piece um and that's kind of
what we recommend it's just like you can
really and you also the one thing is you
don't want to be an echo chamber you
don't want to only listen to it but
that's why I think get to have kind of
separate Twitter accounts that are like
okay here's my AI Twitter account and
then here when I want to have like
worldly polit political views and
understand everything else that's going
on in society here's my other account um
but I think like it is good to kind of
have that targeted here are like the
hundred people I follow that I trust
that are posting great content and
doesn't take a long time you just need
to go into these communities for a few
minutes every day but as long as you're
doing that repetition you're going to
learn stuff guaranteed for some people
it's not immediately obvious the return
on that learning but just to give people
something more tangible every company
will want to keep those who are on the
top right like on the top of new things
new stuff Cutting Edge as you said like
the guy who's making presentation for
his management team the CH the chances
of him getting fired are very low right
because like he's teaching others and
he's benefiting the company more than
he's being paid for which is always a
good strategy but especially with like
new technologies that AI you know if if
the company is going to make a layoffs
whatever you're probably going to be the
last person on the list because you know
you're the AI guy like they don't want
to lose you like you always bring the
good stuff so you know everybody who's
watching this who has a job if you learn
Ai and if you you know teach your bosses
and the higher up you go the better
obviously like you you future prove
yourself and you the chances of you
keeping that job or maybe even getting
promoted are much higher for sure so I
highly recommend like testing things out
seeing like like this is my one
recommendation like I think you like
these like kind of acual bits I have
right here is like my notebook and
basically every day as I'm going through
my workday I literally just write down
things that I think can be automated
okay so I have like literally like pages
of like hey I did this workflow I think
I can automate this what I do is I use
this as now like a brain for me to think
about when I'm looking at different
tools can I use that tool to automate
this process and it's such an easy thing
for anyone to do it doesn't cost
anything um and I really recommend like
people like leaning in a bit there
because it will like at least now you
have a reference point when you're
looking at tools and looking at the
latest and greatest stuff of what's
possible and what what you actually can
start kind of automating okay so on that
note like what would you say are some of
the like most obvious like painfully
obvious things you see other people do
that could be like automated so easily
research I think is the number one like
I I have friends and and I love them
they're my good friends if you're
watching this but like they I have
friends who are like Financial research
and equity research and I think it's
great it's a really really cool job um
but it's one of these things that is
100% becoming automated it's you know a
computer or an agent or what whatever
you want to call a piece of software is
going to be able to research more than
you can cons this the information better
now with it referencing so like a lot of
the work over the past couple months has
been like how do we actually Source
where this data is coming from
accurately the second you hit off all
these criteria of like you know being
accurate being fast being able to
consume more data you you can't even
compete in that space um same thing with
like the you know the agent I was
talking about for uh researching people
I'm meeting with you are setting
yourself back by not leaning in and
using this technology and it's like it's
so obvious to me that like these are
better than it's not even like human
can't do these jobs it's just like
there's no reason a human should be so
if my assistant had to spend four hours
a day working on this that's four hours
she would be able to be helping with a
bunch of other things that that she has
on her agenda so I'm very big on like
figure it out I think like trying figure
out what aspects are the most
automatable in your life um and then
kind of you know work backwards from
there when it goes into the digital
realm people have these weird feelings
right but you can take examples with
animals like you know horses are much
better at certain things than people or
there's like some breed of pigs that can
smell like some specific potatoes yeah
yeah 10t underground and like no same
person would try to compete with those
pigs but for some reason when it's like
AI agent it's like whoa whoa whoa no
like that's my identity like I was doing
that task like I think you might just
gave me a tweet for today I like that I
I never really thought about it from
like the animal perspective but you're
100% right you know humans should be
focused on human tasks like I you know
one of the I say it's like right on my
LinkedIn it's like don't worry your job
is safe because the reality is it's like
there are inherently human parts of
every role and I think the biggest thing
you know if you're testing this
technology is lean in and try and figure
out which parts of these tasks are the
most human and that's a that's the
biggest when you can be making this is
something that I think most people in
the as space probably don't agree upon
at least I don't see it being talked
about but my theory is that at least in
the next 5 to 10 years before you know
some major improvements to the LM
architecture made
the by far winner combo is going to be
like human does the creative and like
you know directional stuff while llms do
the repetitive and you know llm stuff
good task but right now people always
think like okay there's like two
extremes right llms have stopped they're
not going to improve blah blah blah and
then it's like you know oh AGI ASI super
intelligence is coming in six months
right like we have we all saw these
videos at the start of the year and now
it's like six months later and like
where's where is the AGI where's the ASI
right so I think the actual reality for
the next 5 years is going to be that
humans do the things that LMS cannot do
like reasoning like creativity the stuff
that LM suck at and then just llms do
everything else and people who Embrace
this new way of working will just Crush
everybody else I couldn't G more there's
one company that we're good friends with
so payman I don't know if you've seen
them um you should get so tyin from
payman he's awesome he's built like an
agent that can pay humans to do tasks
that the agent can't complete so is like
happening like so like think about a u
let's give an example of like like
writing a blog post or liking a tweet or
retweeting something you can literally
the agent could say hey we want to
create uh marketing Buzz for this new
blog post and it will go and say for
every human that likes it will pay them
50 cents for every person who retweets
it will pay them a dollar and it will
verify that the human's done it it'll go
through the whole Loop and then it will
go back and pay the human and this is
the type of thing where like it's
happening like this is like obviously
they're on the cutting Edge I think it's
an awesome company we're we we're very
bullish on them working on some really
cool stuff together um but it's like
this is going to happen inherently these
agents are they have limitations like
you'd be maybe I don't know you could
talk to other people who might say
they're you know are infinite and can do
anything but we don't have AGI yet so
right now like there are limitations
with these things um I think it's really
really cool to see that people are
thinking about okay how do we actually
leverage like keep humans in a loop in a
necessary way like I think that's a huge
huge huge win yeah and I think once the
llms um actually learn you know what
llms are good at and I mean maybe you
can do it with a system prompt but you
know like basically we need the agents
to have like a deep understanding of
these are the type of tasks agents are
great at don't bother human these are
the tasks that agents suck at you know
pink the human as you said proactively
and like I'm stuck here can you help me
out and then the human like unstuck the
agent and I mean it will really be what
you just said right all of us even the
people who don't have companies will be
basically CEOs or managers of large
teams of agents and you know you will
just be overseeing different operations
and I think that's like really
empowering Vision dve we we'll send I'll
send you a bunch of links to kind of put
down below so we can get everybody up to
speed as quickly as possible um and
besides that yeah it's really really
great getting to chat don't let the AI
Revolution happen without you if you
join the new Society in the next 48
hours you'll get a personalized plan to
start your own AI business and you can
do this even if you are a complete
beginner because we help you every step
of the way but don't take my word for it
just look at all these reviews from our
members so click the link below and join
the new Society
Browse More Related Video
Chatbot or AI Agent Setting up crewai framework for scaling tasks
How you should think about AI Agents this 2024. (Early Mover Advantage)
World Changing: Data Science and AI | Fred Blackburn | TEDxWakeForestU
OpenAI'S "SECRET MODEL" Just LEAKED! (GPT-5 Release Date, Agents And More)
Episode 1- Efficient LLM training with Unsloth.ai Co-Founder
STUNNING Step for Autonomous AI Agents PLUS OpenAI Defense Against JAILBROKEN Agents
5.0 / 5 (0 votes)