0:00
Here we go. I finally installed LMIMOS
0:02
on my computer and I'm going to launch
0:04
it right away to see what it does. So,
0:06
let's go. I finished installing and I
0:08
kept the box checked that starts LMIM
0:10
directly. So, after installation
0:13
finishes, here we go. It's starting up.
0:15
Let's see what happens. And here we are.
0:18
I have LMIM, an artificial intelligence
0:21
on my computer. It asks for my name. So,
0:24
my name, for example, here it is. I'll
0:28
And I can even give a name to my
0:30
assistant. I can leave LM or I can put
0:33
my website name, yb.digital.
0:36
You can choose the response tone. There
0:38
are many options. I want it to be
0:40
professional, not too creative, just
0:43
professional and efficient. We continue.
0:46
Security notifications. These are
0:49
autonomous actions. LM can execute shell
0:52
commands, can read and write files, and
0:54
can send messages on your behalf. Of
0:57
course, that's why we're installing it,
0:59
but you need to be aware. There's no
1:01
guarantee. The software is provided as
1:03
is. The usual disclaimers. If you lose
1:06
data or if there are actions different
1:08
from what you intended. Of course, you
1:10
need to be careful. AI can be a powerful
1:13
tool and you should always supervise
1:15
critical actions, especially code
1:18
execution and file deletion before
1:20
confirming. And of course, don't share
1:22
your API keys. If you connect your
1:25
software elsewhere, never share your
1:27
keys or session files and you are the
1:29
sole responsible for securing your local
1:31
instance. I understand that LMIM is a
1:34
powerful tool and I accept
1:36
responsibility for its actions. I
1:38
continue. And here I am on LMIMOS, an
1:42
autonomous AI workspace that's on our
1:44
local computer and runs without
1:46
internet. To prove it really works
1:49
without internet, let's simply
1:50
disconnect internet. Here, I disconnect
1:53
internet. No more internet. Let's see
1:56
what happens. You can have conversations
1:58
and web searches. Now I've disconnected
2:00
internet so web search won't work but we
2:02
can still have conversations. It can
2:05
manage calendars, talk via WhatsApp,
2:08
reply to messages, reply to prospects,
2:11
manage clients, create appointments,
2:13
create projects, automatically respond
2:16
to demons, background tasks running on
2:18
the computer, and even do email
2:22
Here I have the AI assistant on our
2:24
desktop. This is where we talk to LM IM
2:27
running locally on our computer.
2:30
No cloud needed. For example, it can
2:32
schedule an appointment for tomorrow at
2:34
3:00 p.m., create it, and send a
2:37
confirmation via WhatsApp.
2:39
We also have web search included. It
2:42
will search results on the web and think
2:44
based on those results all within our
2:46
chat. We won't do that now since we're
2:48
offline. Then we have the visual
2:50
calendar. If you've used your personal
2:53
assistant to manage your calendars, for
2:55
example, inside your organization, if
2:58
you deploy LM to all your employees, you
3:01
can visually manage calendars and it can
3:04
handle appointment scheduling, create
3:06
appointment with clients, manage travel,
3:11
And then we have demons. Demon services
3:13
for WhatsApp, email, slide, Discord. You
3:19
I advise contacting LM IM, especially
3:21
for businesses. It can connect to almost
3:25
Demons are background tasks. LM IM can
3:29
see what's happening in the background.
3:31
When events happen, for example, you
3:33
receive a message on WhatsApp or
3:35
Telegram, LM can detect that message and
3:38
perform actions based on it.
3:40
We also have live view to see what's
3:42
happening, what it's responding to, etc.
3:44
And of course, we have logs. You can ask
3:47
LM to create a project in the chat and
3:49
watch the live space appear. It can
3:51
create and execute code on its own.
3:54
Then the configuration setup, you can
3:57
switch from local to cloud. You can also
3:59
configure email to send actual emails.
4:02
You'll need to configure your real email
4:04
address with an email server to do that.
4:06
So, here we are inside LMIM. Let's say
4:09
hello. It only speaks English or
4:12
Here we'll speak English. I'll ask a
4:14
question. For example, I work on SAP ERP
4:18
and I'll ask how to create a material in
4:22
Let's see what it does. I ask a
4:24
question, it thinks. Again, I'm offline.
4:27
I disconnected internet. Let's see if it
4:31
My computer is pretty slow, so don't
4:34
To have AI running locally on your
4:36
computer, you usually need a beast. A
4:39
good processor, GPU, CPU, which I don't
4:43
have. So, it's not going to be fast.
4:46
Normally, we use AI in the cloud running
4:49
on server farms that are very powerful
4:51
and can handle heavy workloads like
4:53
asking a complicated question.
4:56
I hope this works. I honestly don't
4:59
This is my first conversation with LMIM.
5:02
I'm offline. Everything is running on my
5:05
computer. Look, my PC resources are
5:07
being used. CPU is maxed out. Memory is
5:11
almost full. Let's look at the details.
5:13
You can see the task manager. It's not
5:16
using my GPU much, but my CPU is maxed
5:20
I have a small CPU at 1.45 GHz. 11th gen
5:25
Intel, 2.8 GHz, and it's at 89% usage by
5:30
LMIM to generate the response.
5:33
I didn't ask a simple question. I asked
5:36
a complex one on purpose to test it.
5:40
I have 32 GB of RAM on my computer and
5:43
23 GB are used. 8 GB left. It's okay,
5:48
but my computer is being pushed to its
5:50
limits. A small laptop, not extremely
5:53
powerful. Of course, the more powerful
5:55
your computer, especially with dedicated
5:58
AI processors, the faster it will run.
6:01
Oh, apparently the response is arriving.
6:04
Okay, it understood. Not exactly what I
6:07
expected. Let's just say thank you to
6:09
get the response. It's not specifically
6:11
designed for this, especially offline,
6:14
since I'm asking a very complicated
6:16
expert question about SAP.
6:19
I wonder if putting it online would
6:20
change anything. Let's try. I'll
6:23
reconnect to internet. Now we're online.
6:26
Let's see what happens.
6:28
The CPU is still maxed out. Memory near
6:31
full. I'll ask a simple question.
6:33
The point of having local AI isn't about
6:36
speed. Our local computers are nowhere
6:38
near as powerful as server farms. The
6:41
point is that data stays on your
6:42
computer. Nothing is sent to the
6:44
internet and it works offline.
6:47
Ah, here's the response arriving. It
6:49
just needed a little waiting time.
6:51
How to create a material in SAP? It gave
6:54
me a complete article. First, of course,
6:57
you need to choose the material type.
6:59
So, it was a very complicated question.
7:01
I did that on purpose. Anyway, you can
7:04
get responses locally with a free AI, so
7:06
I invite you to download it. Link in
7:08
description. Try it yourself and contact
7:11
me if you want to install it for your
7:13
business and do more advanced things.
7:16
That's it. Download LMIM.
