00:00:00
as with most weeks lately it has been a
00:00:02
really busy week in the world of AI and
00:00:04
I don't want to waste your time so let's
00:00:06
go ahead and break it all down right
00:00:08
[Music]
00:00:10
now starting with the fact that open AI
00:00:12
is finally bringing their advanced voice
00:00:15
mode to the web now we've already had
00:00:17
advanced voice mode available on the
00:00:19
mobile app for a little bit now they
00:00:21
recently rolled it out on the desktop
00:00:23
app well now it is going to be available
00:00:26
inside of the browser app this is based
00:00:28
on an announcement from Kevin wheel here
00:00:30
on X he says rolling out to chat GPT
00:00:33
paid users this week advanced voice mode
00:00:35
on the web down here at the bottom of
00:00:37
his post he says give it a try at chat
00:00:39
gp.com if you're a plus Enterprise teams
00:00:42
or Edge you subscriber we'll look to
00:00:44
roll it out to free users in the coming
00:00:45
weeks let me know what you think that
00:00:47
seems to imply that it's already in
00:00:49
there but when I head over to my own
00:00:51
chat GPT account I'm not seeing any
00:00:53
options here to use the new advanced
00:00:55
voice mode yet so it seems to be a roll
00:00:58
out maybe some people have it already
00:01:00
and it is still rolling out but I don't
00:01:01
have it in my browser account yet to
00:01:03
show off but something that did
00:01:04
definitely roll out this week was a new
00:01:07
update to GPT 40 according to open AI
00:01:10
the model's creative writing ability has
00:01:12
leveled up it's more natural engaging
00:01:14
and it's tailored writing to improve
00:01:16
relevance and readability it also
00:01:18
apparently works better with uploaded
00:01:19
files providing deeper insights and more
00:01:22
thorough responses again if you head
00:01:24
over to chat GPT if you're using the GPT
00:01:27
40 model you'll be using the latest
00:01:31
version of the 40 model I've played
00:01:33
around with it a little bit I find the
00:01:34
updates to be very subtle but I also use
00:01:37
claw a little bit more than I use chat
00:01:39
GPT so I'm not sure I would totally know
00:01:41
the difference at the moment since we're
00:01:43
talking about chat GPT remember back
00:01:45
when they showed off the advanced voice
00:01:47
feature in that demo video well they
00:01:49
showed it off with the ability to
00:01:52
actually get live video and have chat
00:01:55
GPT actually see what you're looking at
00:01:57
to be able to inform its responses well
00:01:59
that feat never really rolled out with
00:02:01
the advanced voice mode like we thought
00:02:03
it would well it looks like it's going
00:02:04
to be coming very soon I found this out
00:02:07
from Rowan Chong here who spotted it
00:02:09
from Android authority but it looks like
00:02:11
the code for those features was actually
00:02:13
discovered in the latest beta version of
00:02:16
GPT they discovered in the code live
00:02:18
camera functionality real-time
00:02:20
processing voice mode integration and
00:02:22
visual recognition capabilities I'll
00:02:24
make sure this tweet thread is linked up
00:02:27
in the description below but you can
00:02:28
actually see screenshots of the code
00:02:30
here that make it look like it's going
00:02:32
to be coming pretty soon moving on to
00:02:34
anthropic news they just rolled out a
00:02:36
new feature if you use clad you can now
00:02:39
directly add documents from Google Drive
00:02:41
if I head on over to my clad account
00:02:43
here you can see there's a new little
00:02:45
Google Drive button here if I click this
00:02:47
button here it shows me my recent files
00:02:50
inside of Google Drive and I can also do
00:02:52
a search for specific files in Google
00:02:54
Drive and whatever file you pull in will
00:02:55
work the same way as if you were to just
00:02:57
drag and drop a file into Cloud Google's
00:02:59
Gemini got an update this week it can
00:03:02
now actually remember things so chat GPT
00:03:05
has had a memory feature for a little
00:03:07
while now now Google's Gemini has it as
00:03:09
well when you're logged into Gemini here
00:03:11
you can actually come down to your
00:03:13
little settings Cog wheel and you'll see
00:03:15
a button that says saved info if you
00:03:16
click in there you can actually see all
00:03:18
of the info that's saved you can add
00:03:20
info or you can ask Gemini to remember
00:03:22
something while you're actually chatting
00:03:24
with it at the moment it says this
00:03:25
feature is only available in English but
00:03:27
here are some examples that Google gave
00:03:29
of the types of things you might want to
00:03:30
save use Simple language and avoid
00:03:32
jargon I'm a vegetarian so don't suggest
00:03:34
recipes with meat after responding
00:03:36
include a Spanish translation when trip
00:03:38
planning include the cost per day I can
00:03:40
only write code in JavaScript I prefer
00:03:42
short concise responses you can also
00:03:44
give it info about yourself like your
00:03:46
name and your career and where you live
00:03:48
and things like that so it can be better
00:03:51
informed when it's actually giving you
00:03:53
its responses and since we're talking
00:03:55
about Google I want to point out
00:03:56
something really cool that YouTube is
00:03:57
doing I've mentioned in the past but it
00:03:59
looks like it's actually finally
00:04:01
starting to roll out they're now going
00:04:03
to start doing automatic dubbing on
00:04:05
videos into multiple languages and the
00:04:07
YouTube creators don't have to do
00:04:09
anything extra so for someone like me
00:04:11
all I need to do is upload my video to
00:04:13
YouTube like normal and it will
00:04:15
automatically create a translated
00:04:18
version of it in Spanish Portuguese
00:04:20
German French Italian Hindi Indonesian
00:04:23
and Japanese that's super super exciting
00:04:26
to me and it should be for any content
00:04:27
creator that's publishing on YouTube
00:04:29
because it means like every single piece
00:04:31
of content you make is going to be
00:04:33
available to an audience that's like 10
00:04:35
times bigger than what you're already
00:04:36
getting that's going to be huge there's
00:04:38
a new model out from a company called
00:04:40
Deep seek called Deep seek R1 light
00:04:42
preview which is a model that's going
00:04:44
head-to-head with open AI 01 now if you
00:04:47
remember the open AI 01 model is the
00:04:49
model that sort of thinks things through
00:04:52
before it gives you a response and is
00:04:54
much better at sort of logic and
00:04:57
questions that require some detail
00:04:59
thought behind them and up until now
00:05:02
open AI 01 model was really the only
00:05:04
model out there that had that sort of
00:05:06
focus well now there's a model out of
00:05:08
this deep seat Company from China that
00:05:10
can seemingly act in the same way and we
00:05:13
can see on these benchmarks here that in
00:05:15
things like math and code it actually
00:05:18
outperforms the 01 preview model now
00:05:20
keep in mind this 01 preview model isn't
00:05:23
the full 01 model it's like a a
00:05:25
checkpoint of the model we actually
00:05:27
haven't seen the full open ai1 model in
00:05:31
fact rumor has it it's going to be
00:05:32
coming out within the next few weeks but
00:05:34
again that's just a rumor but this deep
00:05:36
seek model that is beating 01 preview in
00:05:39
a few of these benchmarks here it's not
00:05:40
really a comparison head-to-head with
00:05:42
the full 01 model that we're going to be
00:05:44
seeing soon now the French AI company
00:05:47
mistol also has some updates to their
00:05:50
chatbot called Le chat it now has search
00:05:53
capabilities Vision capabilities
00:05:55
ideation and coding and you can use it
00:05:58
all for free so if I come up here to
00:06:00
leat we'll go ahead and sign in and very
00:06:03
similarly to chat GPT we've got the
00:06:06
option to use canvas we've got the
00:06:08
option to search the web and we've got
00:06:10
the option to generate images theet now
00:06:13
has some new features web search with
00:06:15
citations canvas for ideation and inline
00:06:18
editing state-of-the-art document and
00:06:20
image understanding image generation
00:06:22
powered by black Force Labs flux Pro and
00:06:24
we got a little breakdown here comparing
00:06:27
chat GPT perplexity Claude and showing
00:06:30
that most of the features that are paid
00:06:32
in these other platforms are actually
00:06:34
free here in theat give me at least 10
00:06:37
YouTube channels to follow to stay up to
00:06:39
date with AI we've got the pie coach
00:06:40
two-minute papers some bum Conrad
00:06:43
cording deep learning AI Robert Miles AI
00:06:45
explain yanic kilter vandaline valov and
00:06:48
AI in games some of these are actually
00:06:50
channels I hadn't heard of yet so more
00:06:52
for me to check out if you're on a
00:06:54
budget and you don't want to spend any
00:06:55
money on a chatbot like chat GPT or CL
00:06:59
but you want a lot of the same features
00:07:01
give mol's lechat version a try cuz it
00:07:03
does a lot of the same stuff that the
00:07:05
paid versions of those do but for free
00:07:07
and just for fun let's see how it does
00:07:09
with image generation generate an image
00:07:11
of a cat eating a taco and that's what
00:07:13
we get it took about 18 seconds to
00:07:15
generate and again it's using flux one
00:07:18
pro which if you've used the AI
00:07:20
generator over in Gro it's the same one
00:07:22
that that one uses but again it's free
00:07:24
to use tools like chat GPT and CLA have
00:07:26
been absolutely amazing for businesses
00:07:29
for brainstorming ideas for marketing
00:07:32
and branding and business strategy but
00:07:35
it's not always super obvious what
00:07:36
prompts are going to give you the
00:07:38
absolute best results for the goals that
00:07:40
you have in your business and that's why
00:07:42
for this video I partnered with HubSpot
00:07:44
to share their AI prompt library now
00:07:47
this prompt library was custom curated
00:07:49
by the CMO and the SVP of HubSpot Kip
00:07:52
and Kieran also the host of marketing
00:07:54
Against the Grain now as somebody who
00:07:56
has a podcast on the HubSpot Network I
00:07:58
know these guys personally and they know
00:08:00
their stuff they know how to craft
00:08:02
amazing prompts around things like
00:08:04
marketing and sales strategy branding
00:08:06
and advertising business strategy and
00:08:08
Analysis pitching and investing website
00:08:11
strategy and so much more want to write
00:08:13
headlines like David Ogie or sell your
00:08:15
products using a proven storytelling
00:08:17
framework all of these prompts are
00:08:19
available to you completely for free and
00:08:21
they're designed so you don't need to be
00:08:23
a prompt engineer to get amazing results
00:08:26
from these AI tools I've gone ahead and
00:08:28
put the link to this 100% free resource
00:08:31
down in the description below you're
00:08:33
going to want to check this out and
00:08:34
thank you so much to HubSpot for
00:08:36
sponsoring this video there was a lot of
00:08:38
news out of Microsoft this week because
00:08:39
Microsoft held their annual ignite event
00:08:42
an event where they typically make a ton
00:08:44
of announcements I actually don't know
00:08:46
if this announcement came out of the
00:08:48
ignite event but Microsoft signed a deal
00:08:51
with Harper Collins which is a book
00:08:53
publishing company to essentially get
00:08:56
permission to train AI models on the
00:08:59
book books from Harper Collins now
00:09:01
Harper Collins is going to get
00:09:02
permission from each individual author
00:09:04
as it sounds like but with all the
00:09:06
lawsuits that have been happening over
00:09:09
news sources and authors and people like
00:09:12
that actually suing these AI companies
00:09:14
it's starting to sound like the big
00:09:15
companies are starting to ask for
00:09:16
permission a little bit more frequently
00:09:18
now if you remember back at Microsoft
00:09:20
build earlier this year they announced a
00:09:22
feature called recall which was like
00:09:24
your web history but for your entire
00:09:27
computer so you can go back and remember
00:09:29
what you were doing in certain apps
00:09:31
throughout the day or you know several
00:09:33
days ago maybe you were doing something
00:09:34
in Photoshop but you wanted to go back
00:09:36
to that exact moment you can actually go
00:09:38
back in history on your computer well
00:09:40
there was a lot of concerns when they
00:09:42
announced that they went back to the
00:09:44
drawing board and really really dialed
00:09:46
in like the security and just overall
00:09:49
functionality of it to make people a
00:09:51
little more comfortable with it well as
00:09:52
of today November 22nd co-pilot PCS with
00:09:56
the Snapdragon chip in them are now
00:09:58
going to have the ability to access this
00:10:00
recall feature finally now in order to
00:10:03
access this you need to have the windows
00:10:05
11 Insider preview build and of course
00:10:07
you need to be using one of the newer
00:10:09
co-pilot PCS here's what the new feature
00:10:12
looks like you have this little scroll
00:10:14
window across the top of your recall
00:10:17
page here and you can scroll back to any
00:10:19
moment you can see you've got today
00:10:21
October 21st October 20th October 19th
00:10:23
so you can go back to any day in time
00:10:26
and remember what you were doing back on
00:10:28
that day day and actually get
00:10:30
screenshots of that moment that you're
00:10:32
looking for you can also type to search
00:10:35
your history as well there's also a
00:10:37
little icon that will pop up down in
00:10:39
your system tray here that will allow
00:10:41
you to pause the recording so if there's
00:10:44
websites you're looking at that for
00:10:46
whatever reason you don't want recall to
00:10:49
remember that you looking at that
00:10:51
website you can always pause it and make
00:10:53
sure it's not remembering that specific
00:10:56
moment that you're looking and if you do
00:10:58
happen to record moments on your
00:11:00
computer that you wish weren't recorded
00:11:02
maybe you were doing some gift shopping
00:11:04
or something like that you can go back
00:11:05
through and actually delete those
00:11:08
moments from your recall history so that
00:11:10
they're completely removed there's also
00:11:12
security measures in place so to even
00:11:15
open the recall window and see what's in
00:11:17
there it's going to double check and
00:11:19
make sure it's you through whatever
00:11:21
security measures you have set up on
00:11:23
your PC and if you really really don't
00:11:25
want to use it you can turn it off
00:11:26
completely and not use recall at all on
00:11:29
your computer so if you're concerned
00:11:31
about buying a co-pilot plus PC and
00:11:34
you're considering not getting one cuz
00:11:36
you don't like the recall feature it is
00:11:37
possible to disable it completely
00:11:39
there's also a new feature called click
00:11:41
Todo where you can click on an image and
00:11:43
it'll give you options like visualize
00:11:45
search with Bing blur the background of
00:11:47
the photo erase objects in the photo
00:11:50
remove the background with paint and
00:11:51
things like that so when you click on an
00:11:54
image you have all sorts of new AI
00:11:56
related options and again these features
00:11:58
roll out the today this video is going
00:12:00
live November 22nd so if you already
00:12:02
have a co-pilot PC with one of the
00:12:04
Snapdragon 8 or better chips in it you
00:12:07
should be able to upgrade to the new
00:12:09
Insider preview build of Windows and
00:12:11
start using these features another thing
00:12:13
Microsoft announced this week is that
00:12:14
it's going to let you clone your voice
00:12:16
on teams meetings this is really cool
00:12:19
for if you're speaking to somebody in a
00:12:21
different language you'll be able to
00:12:23
speak in your native language and it
00:12:25
will actually translate it and localize
00:12:28
it to whoever you're talking to and
00:12:29
it'll still sound like your
00:12:36
voice review the progress of the project
00:12:39
can you please tell us how the tasks we
00:12:41
should take are going so I'm super
00:12:42
excited to see that come out inside of
00:12:44
Microsoft teams I would love to see that
00:12:47
feature roll out into things like zoom
00:12:49
and Google meet as well soon 11 Labs
00:12:52
rolled out a new feature this week to
00:12:53
build conversational AI agents so if I
00:12:56
log into my 11 Labs account you can
00:12:59
there's a new button over here called
00:13:00
conversational AI let's go ahead and try
00:13:03
a call and see what it sounds like I'll
00:13:04
use the video game character here let's
00:13:07
just see how this goes greetings
00:13:08
traveler how can I assist you on your
00:13:10
quest what is your
00:13:12
name greetings traveler I am Callum a
00:13:16
wizard here to help you on your quest to
00:13:18
collect the 11 gemstones scattered
00:13:21
across Labs onia I can help you find
00:13:24
them but I can't do it all for you what
00:13:26
can I do to help you today what are
00:13:28
these gems
00:13:30
for ah the 11 gemstones they are said to
00:13:34
hold immense power capable of granting
00:13:37
wishes and even altering the very fabric
00:13:40
of reality Legend has it that they were
00:13:43
scattered okay that's fascinating let's
00:13:44
go ahead and try this myself now so I
00:13:46
actually just told it to create that
00:13:48
character for me and you can see it
00:13:50
dumped it into this AI agents list over
00:13:53
here and I can see what it looks like to
00:13:55
actually build an agent now so the agent
00:13:58
language we want English we can have it
00:14:00
set on the very first message when
00:14:02
somebody starts talking to it there's a
00:14:03
system prompt with the details wizard
00:14:06
name Etc which large language model do
00:14:08
you want to use and it's got quite a bit
00:14:10
of options Gemini 1.5 flash is the
00:14:13
default cuz apparently it's the fastest
00:14:15
model in here but you've also got GPT 40
00:14:18
clae 3.5 Sonet and even a custom llm
00:14:21
which looks like you can set up through
00:14:23
using an API key but I'm not going to go
00:14:25
there for this one let's just leave it
00:14:26
on the default you can set the
00:14:27
temperature and then it's got a base so
00:14:29
we can add data to this knowledge base
00:14:32
so if I wanted to make a AI newsbot I
00:14:35
could take the transcriptions from all
00:14:37
of my old videos throw them in this
00:14:38
knowledge base and people can ask
00:14:40
questions about all of my old videos you
00:14:43
can add tools which apparently uses web
00:14:46
hooks here you can add Secrets okay so
00:14:48
this is probably for adding you know API
00:14:51
keys and things like that I'm guessing
00:14:52
you can select the voice it used Callum
00:14:55
but I have my own Matt wolf voice that I
00:14:56
trained in 11 lab so I can make it sound
00:14:58
like me if if I want we've got the
00:15:00
normal settings you get from 11 Labs all
00:15:02
sorts of really cool features to go
00:15:05
ahead and build out a bot that has the
00:15:08
information and the voice of whatever
00:15:10
you want it to be so that's that's kind
00:15:12
of cool and fun I can see the main use
00:15:14
case probably being like customer
00:15:15
support bots on websites and things like
00:15:17
that but this is something I definitely
00:15:19
want to do a deeper dive in and try to
00:15:20
build some characters with later so
00:15:23
maybe that'll be a video we circle
00:15:24
around to in the future 11 Labs also
00:15:26
rolled out the ability to Auto
00:15:28
regenerate so if you're using their long
00:15:30
form Text Editor to generate audio it
00:15:33
looks like if something doesn't quite
00:15:34
sound right you can highlight that
00:15:36
selection and tell it to regenerate it
00:15:38
and it will do it again and ideally it
00:15:40
gets the sort of tone and inflections
00:15:41
right the second time around sunno got a
00:15:43
huge upgrade this week as well with
00:15:45
their new V4 and I actually got Early
00:15:49
Access to this and was able to play with
00:15:50
it a little bit before they released it
00:15:52
and honestly it is pretty dang
00:15:54
impressive it actually generates music
00:15:57
that i' want to listen to and a few of
00:15:59
these songs I've actually gone back and
00:16:00
listened to multiple times one of the
00:16:02
things I was having fun doing and I hope
00:16:04
I don't get in trouble for this but I
00:16:06
was taking lyrics from existing songs
00:16:08
that I like and changing the style like
00:16:11
damn it from ble 1882 this one I told to
00:16:13
make it electronic poppy dance version
00:16:16
and it reminded me of Owl City so that's
00:16:17
why I named it that but listen to
00:16:21
this it's all right to tell me what you
00:16:25
think about me I won't try to argue or
00:16:30
hold it against you I know that you're
00:16:34
leaving you must have your reasons the
00:16:38
season is calling your pictures are
00:16:42
falling
00:16:44
[Music]
00:16:47
down I actually think it's good I
00:16:49
actually like listening to that version
00:16:51
of this song and then I did it like a
00:16:53
poppy dub style and then I did it like a
00:16:55
country version like it's all right
00:17:00
tell
00:17:01
me what you
00:17:04
think about me here's why subscribed to
00:17:08
Matt wolf song that's a punk
00:17:10
[Music]
00:17:26
version Tech Whispers through the
00:17:28
digital trees AI news brings you to your
00:17:32
knees finally this was actually a song
00:17:35
that I did witho version 3 but if you do
00:17:38
a song with sunno version 3 and you want
00:17:41
to sort of bring that song up to date
00:17:43
with the new version 4 you can go to
00:17:45
your old songs click on these little
00:17:46
three dots here click on create over
00:17:49
here and there's actually a remaster and
00:17:51
it'll convert the song into a V4 song
00:17:54
and actually make it even more
00:17:55
impressive sounding so sonov V4 has been
00:17:58
really really fun to play with we got a
00:17:59
pretty big update in the world of AI art
00:18:02
as well black forest Labs released flux
00:18:05
one tools which actually gives flux one
00:18:08
all sorts of new abilities we're getting
00:18:10
closer and closer to being able to do
00:18:12
with flux one what we have always been
00:18:14
able to do with stable diffusion AI art
00:18:16
things like inpainting and out painting
00:18:18
we can see this picture here where he's
00:18:20
starting with this gray suit they mask
00:18:22
out the suit and then he turns it into
00:18:24
like a jean jacket here or this sign
00:18:26
that says burgers and beers mask out the
00:18:28
burgers change it to coffee and beers
00:18:30
and you can see it maintains that look
00:18:32
that style and then here's an example of
00:18:34
out painting where they gave it a
00:18:36
picture of just an eyeball and then it
00:18:37
out painted the rest of the face they
00:18:40
also have what they're calling
00:18:41
structural conditioning with flux with
00:18:43
canny and depth models now if you're
00:18:45
familiar with control Nets with stable
00:18:47
diffusion this is essentially the same
00:18:49
thing as the control Nets you've got a
00:18:51
depth map so you can give it an image
00:18:53
and it tries to estimate the depth in
00:18:56
that image like you see here in this
00:18:58
gray image and then you give it a prompt
00:19:00
it will create new images trying to
00:19:02
model this same sort of depth here and
00:19:05
then this is the cany model here where
00:19:07
you upload an image it sort of makes
00:19:09
like a traced outline of the image and
00:19:11
then your new images follow that same
00:19:13
sort of traced outline but in a
00:19:15
completely new style they also added
00:19:17
image variation and restyling with flux
00:19:20
so here's some examples here where you
00:19:21
combind an input image with a prompt
00:19:24
like they uploaded this duck here and
00:19:26
then said fine glass sculpture on a
00:19:27
shelf and it made this image of the same
00:19:30
duck but as a glass sculpture or a
00:19:32
simple 3D model or actual duck on a desk
00:19:35
or epic anime art and all of it modeled
00:19:37
the look of this duck but in the new
00:19:39
style that was prompted here's a guy
00:19:41
taking a selfie like out on a hike or
00:19:43
something flat Vector simple material
00:19:45
art large hand puppet with brown fur in
00:19:47
the forest pen and ink sketch and
00:19:50
detailed anime artwork all from this
00:19:52
original image and then here's this one
00:19:54
with bread and looks like milk and
00:19:56
cheese and they were able to make like
00:19:57
different variations in angles of the
00:19:59
same image and all of this is available
00:20:01
inside of the Black Forest Labs API so
00:20:04
tools that leverage Black Forest Labs
00:20:06
will'll probably start rolling out a lot
00:20:08
of these features soon there's a handful
00:20:10
of websites that you can use some of
00:20:12
this there's file. a replicate together
00:20:14
AI free pick and CR AI if I jump over to
00:20:17
file. a the very top says introducing
00:20:20
flux tools so if I click try it now it
00:20:22
looks like all of these tools are
00:20:24
available here as well let's see if I
00:20:26
click on flux one fill I'll upload this
00:20:29
image of my head shot here and then it
00:20:31
asks for a mask URL let's go ahead and
00:20:33
create one and now let's just go ahead
00:20:35
and you know mask off my face see what
00:20:37
it does then use mask and you can see
00:20:39
here's the mask that I just did and
00:20:41
let's run it and see what happens I
00:20:43
accidentally left the prompt up here a
00:20:44
night and shining armor holding a great
00:20:46
shield with fall on it so it turned my
00:20:50
face into a shield with fall on it let's
00:20:52
try the cany model real quick once again
00:20:55
I'll upload my head shot here I'll leave
00:20:57
the prompt as a pink owl and let's just
00:20:59
run it it should basically make a pink
00:21:01
owl that looks like the shape of my head
00:21:03
and I guess that's its interpretation of
00:21:06
a pink owl but really what it did was it
00:21:08
found the outline in the original image
00:21:09
and then it used that for this so if it
00:21:12
doesn't see outline of an owl in my
00:21:14
original image it's going to struggle so
00:21:16
some fun stuff that you can now do with
00:21:17
flux giving it a lot more controlability
00:21:20
the same kind of controlability we have
00:21:22
with something like stable diffusion
00:21:24
Coca-Cola is in the news related to AI
00:21:26
this week and that they created an AI
00:21:29
generated ad that resembles their old
00:21:32
Christmas ads but just generated with AI
00:21:34
this time around here's a Peak at what
00:21:36
that AI generated ad looks like even the
00:21:39
people in the ad were AI generated this
00:21:41
whole thing fully AI generated uh I
00:21:44
don't know the copyright of the music so
00:21:46
I'm getting rid of the audio here but
00:21:48
they put out that ad and well a lot of
00:21:50
the AI people thought it was really cool
00:21:52
that Coca-Cola is putting out ads with
00:21:54
AI and a lot of anti- aai people were
00:21:57
very pissed off that they used AI
00:21:59
instead of hiring people to make an ad
00:22:02
the normal way they basically said that
00:22:04
they wanted to create an ad that found
00:22:06
the sort of nostalgia of the old
00:22:08
Coca-Cola Christmas ads but they wanted
00:22:10
to bring it into the modern era and they
00:22:12
thought using AI was a really good way
00:22:14
to do that since we're on the topic of
00:22:16
AI video The Tool haen now has an IOS
00:22:20
app haen is a tool where you can sort of
00:22:22
train an AI Avatar of yourself and it
00:22:25
can speak like you and look like you and
00:22:27
lipsync like like you and translate and
00:22:30
still sound like you and now a lot of
00:22:32
this stuff you can do with an iOS
00:22:34
version of it now every week before I
00:22:36
make these videos I go on X and I go on
00:22:39
threads and lately I've been going on
00:22:40
Blue Sky and Linkedin as well and I've
00:22:43
been sharing most of the news that I'm
00:22:45
going to share on these videos and I'm
00:22:47
asking people for help like what news am
00:22:49
I missing that's not on this list so
00:22:52
that I can make sure I get full coverage
00:22:54
of what people are talking about in the
00:22:55
world of AI this week sponsorship Guru
00:22:58
pointed out something that I hadn't seen
00:23:01
yet he said Pickle the AI tool that
00:23:03
allows you to join Zoom teams meet with
00:23:05
your avatar and speak from elsewhere I
00:23:08
actually hadn't seen this one yet so I
00:23:10
had to look up like what is this pickle
00:23:12
thing he's talking about and I found
00:23:14
their Twitter account and it's actually
00:23:16
pretty interesting so check out their
00:23:18
demo video here as you can see I'm
00:23:20
currently participating in a zoom call
00:23:22
but what if I tell you that I'm actually
00:23:24
not in front of the camer right now and
00:23:26
this is an AI generated live screen that
00:23:28
in sync with my voice pickle generates
00:23:31
realtime video of talking you based on
00:23:32
your audio inputs which can be streamed
00:23:34
directly to zoom so you can have an AI
00:23:36
Avatar looking like they're on a zoom
00:23:38
meeting the whole time and you could be
00:23:40
in a different room talking into a
00:23:42
microphone and in real time it syncs up
00:23:45
your voice to the Avatar that's actually
00:23:48
on the zoom call so with something like
00:23:50
this you don't have to worry about you
00:23:51
know brushing your hair or putting on
00:23:53
clothes before getting on a zoom call
00:23:55
perplexity has a new feature because
00:23:56
perplexity just keeps on shipping new
00:23:58
feat features this time around you can
00:24:00
actually purchase products straight from
00:24:02
the perplexity app so you can use
00:24:04
perplexity to do research before buying
00:24:06
a product which is something that I
00:24:08
actually use perplexity for quite a bit
00:24:10
and when you find the perfect product
00:24:11
that you want to purchase you don't even
00:24:13
need to go to the e-commerce store to
00:24:15
buy it anymore you could just press a
00:24:16
button and purchase it straight through
00:24:18
perplexity so you can see if I go to
00:24:19
perplexity here and say compare the
00:24:22
available versions of the meta Quest it
00:24:24
gives me a response comparing these
00:24:26
various versions the three the three the
00:24:28
two and the pro but then if I scroll
00:24:30
down a little bit look it's got all of
00:24:32
them available for purchase right here
00:24:35
in the same perplexity app without me
00:24:36
needing to leave and I could simply
00:24:39
click buy with Pro I would then add my
00:24:41
shipping address my billing address and
00:24:43
then check out once you've added these
00:24:45
details once you don't need to keep
00:24:46
adding them in the future it becomes
00:24:48
that much easier to make purchases
00:24:50
Straight From Within perplexity if
00:24:52
you're into AI coding the tool from
00:24:54
versell called VZ just rolled out some
00:24:56
new features you can now create and run
00:24:58
full stacked nextjs and react
00:25:00
applications create multiple files in
00:25:02
one generation link and deploy to
00:25:04
versell projects and use versel project
00:25:07
environment variables I'm not a coder
00:25:09
myself but I do play around with these
00:25:11
AI coding tools a lot I think cursor is
00:25:13
still my main goto but I haven't played
00:25:16
around with these new features I haven't
00:25:17
really compared VZ to what I'm getting
00:25:20
out of cursor too much yet if you have a
00:25:22
rabbit R1 it seems like we finally got
00:25:25
the feature that they were promising
00:25:27
from the very beginning which is the
00:25:29
ability to teach it how to do things on
00:25:31
your computer so it now has a teach mode
00:25:34
so here's a demo video they shared in
00:25:36
this blog post where you can see they
00:25:38
give it a lesson name a description
00:25:40
they're under this teach tab here and
00:25:42
then from that point on it says
00:25:43
recording every action you make on the
00:25:45
website that teach mode can record will
00:25:48
be displayed in the log below so then it
00:25:50
looks like you just kind of go through
00:25:51
and actually do the steps you want it to
00:25:54
do on your computer it watches those
00:25:56
steps and logs them all and then in the
00:25:58
future you can actually run that set of
00:26:01
action straight from your rabbit so you
00:26:03
teach it once inside of the app on your
00:26:05
computer and then once it's learned how
00:26:07
to do it that one time you should be
00:26:09
able to call upon that action from your
00:26:11
rabbit whenever you want and then it
00:26:12
will go through those steps again I do
00:26:14
have a rabbit R1 I literally haven't
00:26:16
even taken out of the box I haven't
00:26:18
tried it yet but now that the feature
00:26:19
that I was actually excited about for
00:26:21
the rabbit R1 is available this is
00:26:23
something I'm going to be exploring for
00:26:24
sure the startup H just raised 220
00:26:27
million
00:26:28
to launch Runner H which is an agentic
00:26:32
AI that they're trying to build it says
00:26:34
Runner H will initially focus on three
00:26:36
specific use cases robotic process
00:26:39
automation quality assurance and
00:26:41
business process Outsourcing I hadn't
00:26:43
really heard of this company H until I
00:26:46
started prepping for this video but this
00:26:49
will be a company that I'm watching
00:26:50
closely and see how this all plays out
00:26:53
and ideally once there's something for
00:26:55
us to play with and actually show off
00:26:56
I'll show it off in one of my future
00:26:58
future videos if you use the brave
00:26:59
browser when you use their search their
00:27:02
search also now gives you recommended
00:27:05
questions to continue on the search so
00:27:07
there's a little screenshot here if we
00:27:09
take a closer look somebody used Brave
00:27:11
search to type in Christopher Nolan
00:27:13
films it brought up a bunch of films it
00:27:15
then looks like it gave a suggested
00:27:17
question of which actors appear most in
00:27:19
his movies and then it gave an answer to
00:27:22
that this is like their search engine
00:27:23
like Google but it's got this followup
00:27:27
questions area in it unlike Google this
00:27:29
week while Elon Musk was playing Diablo
00:27:32
4 and apparently becoming the best
00:27:35
Diablo 4 player in the world somebody
00:27:37
asked him do you think AGI will come
00:27:39
next year here's his full response
00:27:41
thanks to James Bloom for pointing this
00:27:43
out I think AGI will be next year
00:27:45
probably it's not next year it's I say
00:27:47
2026 at the latest for AGI latest I hope
00:27:50
it's nice to us when defines AGI is
00:27:52
smarter than any human I think it's less
00:27:55
than 24 months away from that Google
00:27:57
Deep Mind shared new research this week
00:27:59
called Alpha Cubit and this is AI to
00:28:01
assist Quantum Computing it says our new
00:28:04
AI system accurately identifies errors
00:28:06
inside of quantum computers helping to
00:28:08
make this new technology more reliable
00:28:10
now Quantum Computing is something that
00:28:12
I've been really really fascinated with
00:28:14
I've been studying it pretty much just
00:28:15
as long as I've been studying AI I just
00:28:17
don't feel smart enough to talk about it
00:28:19
too intelligently yet but basically
00:28:21
Quantum Computing can take certain
00:28:23
problems which take a conventional
00:28:25
computer billions of years to solve and
00:28:27
solve solve them in just hours which is
00:28:30
why in my opinion Quantum Computing is
00:28:32
actually way scarier than AI the
00:28:35
implications to cryptography and
00:28:37
internet security and stuff those
00:28:39
implications to me are even scarier than
00:28:41
the implications of AI but that's a
00:28:43
video for a different day but this new
00:28:45
paper lays out a way to make Quantum
00:28:47
Computing even more effective by
00:28:49
spotting errors and making the quantum
00:28:52
computers more reliable it says here
00:28:54
accurately identifying errors is a
00:28:56
critical step toward making quantum
00:28:58
computers capable of Performing long
00:28:59
computations at scale opening the doors
00:29:01
to Scientific breakthroughs and many new
00:29:04
areas of Discovery I'm not going to get
00:29:05
too into the Weeds on this article but
00:29:07
again I will make sure it is shared in
00:29:09
the description so if you want to dive
00:29:10
deeper this article is available for you
00:29:12
to read and that's pretty much all I've
00:29:14
got for you this week I wanted to share
00:29:15
that Leonardo AI a company that yes I am
00:29:18
an adviser for is having a Black Friday
00:29:21
sale next week so if you want to try out
00:29:24
Leonardo pop back over on Black Friday
00:29:26
where they'll be having a special deal
00:29:28
going on I'm actually thinking about
00:29:30
doing another video next week where I
00:29:32
round up a bunch of cool Black Friday
00:29:33
deals that I've came across uh haven't
00:29:35
decided if I'm going to do that or not
00:29:37
but I tend to mostly sit out on Black
00:29:39
Friday and not buy much but I know a lot
00:29:41
of people wait around for Black Friday
00:29:43
to grab tools that they've always wanted
00:29:45
to grab let me know if you think that
00:29:46
would be helpful in the comments and
00:29:48
maybe I'll make that video as well but
00:29:50
that's what I got for you this week if
00:29:51
you like videos like this and you want
00:29:52
to stay looped in on all of the latest
00:29:54
AI news make sure you subscribe to this
00:29:57
channel I share AI News videos AI tool
00:30:00
tutorials I break down cool AI research
00:30:04
and if you're following my shorts feed I
00:30:06
share a lot of other emerging Tech Ai
00:30:08
and otherwise things like cool 3D
00:30:10
printers and drones and flying cars and
00:30:12
just stuff that I find fascinating that
00:30:14
type of stuff I've been posting a lot on
00:30:16
short form video so again if you like
00:30:18
that kind of stuff give this video a
00:30:20
thumbs up and consider subscribing to
00:30:21
this channel it'll make sure more of
00:30:23
that kind of stuff shows up in your feed
00:30:25
and finally if you haven't already check
00:30:26
out future tools. IO where I curate all
00:30:29
the cool AI tools that I come across I
00:30:31
keep the news page up to date on a daily
00:30:34
basis so all the news I don't get around
00:30:35
to sharing these videos they're on this
00:30:37
news page and of course I've got a free
00:30:40
AI newsletter where I'll share just the
00:30:42
coolest AI tools I come across and the
00:30:45
most important AI news and if you sign
00:30:47
up as a special bonus I'll give you free
00:30:49
access to the AI income database a cool
00:30:52
spreadsheet database of various ways to
00:30:54
use these AI tools to earn some side
00:30:57
hustle income that totally free you can
00:30:58
find it all over at futur tools. thank
00:31:01
you once again for tuning in I really
00:31:03
really appreciate you I do record these
00:31:05
videos on Thursday so if more news came
00:31:07
out on Friday that I missed that'll make
00:31:09
it in next week's video and thank you to
00:31:11
HubSpot for sponsoring this video really
00:31:13
really appreciate you guys thanks again
00:31:15
for nning out with me today I'll see you
00:31:16
in the next one bye-bye
