0:06

ah what is this

0:12

ah tight tight tight yeah ah blue yellow

0:16

pink whatever man just keep bringing me

What is smol AI

0:19

that every once in a while there's a

0:20

project that made me feel we're just

0:22

scrapping the surface with generative Ai

0:24

and small AI is definitely one of those

0:25

projects it got more than 8 000 stars on

0:28

GitHub in just five days and it is one

0:30

of the fattest growing open source AI

0:32

purchase so what is small AI it is

0:34

basically an AI that right codes for you

0:36

it's different from asking chat topt to

0:38

write some simple code snip it actually

0:40

built out entire code base for even very

0:43

complicated apps it's almost like having

0:45

the whole engineer team in your pocket

Demo

0:47

and here's quick example I ask gpd4 to

0:50

write a product requirement DOC for a

0:52

classic snake game and then pass on

0:54

those requirements to the small Ai and

0:56

ask it to build this app and boom I just

0:59

got a snake game that is fully

1:00

functional from small AI if I have more

1:03

requirements I can just come back and

1:04

add more stuff and I can ask it to build

1:06

multiple things for example a simple

1:08

calculator apps and it will also

1:10

automatically build a app for me so so

How does smol AI work

1:12

how does this small AI work when looking

1:14

through the code base it's actually very

1:15

straightforward first they pass the

1:17

product requirement into a prompt they

1:19

will generate architecture of the apps

1:21

to what are different files we need to

1:22

create and then it will pass this list

1:24

of files as well as original product

1:26

requirement to another prompt that will

1:28

generate a list of dependencies so that

1:30

GPD can understand what are the

1:32

different functions and variables in

1:34

different files and in the end it will

1:35

run this for Loop where it will pass on

1:37

both the product requirements

1:39

architecture and the dependency to a new

1:41

prompt that will actually write the code

1:43

for each individual file that's pretty

1:45

much it but to be honest at this moment

1:47

it's still quite buggy most of the time

1:49

you will have some bugs or generate some

1:51

code that require modification to

1:53

actually run properly and I'm going to

1:54

show you how can you run small AI

1:56

without those box step by step firstly

1 Setup

1:59

let's set it up we can go to small AI

2:01

GitHub which I post Link in the

2:03

description and your first idea clone

2:05

this Ripple on your local machine and

2:07

then your open folder in your visual

2:08

studio and there should be a DOT

2:10

example.env file you'll rename this to

2:12

VMV exporting your open AI API key here

2:15

and before you actually run small AI

2:16

this is one last step you should do is

2:18

install model modal is basically a cloud

2:21

computing platform where it's gonna run

2:23

small AI once that's done it's pretty

2:25

much all set you can open the terminal

2:27

and run this command line model and prop

2:29

after that you can add your own prompt

2:31

like write a web app for to do but on

2:34

the other side what I would recommend

2:35

you to do is actually create an MD file

2:38

which is basically a documentation and

2:40

put all the product requirements here so

2:42

that it'll be easier for you to update

2:44

the requirement and you can just run

2:45

prd.md if you just start using small AI

2:48

like this quite often you will see some

2:50

arrows and those arrows are meaning due

2:52

to the rate limit of open AI I think

2:55

that's because when small AI is running

2:57

it will call open AI API multiple times

2:59

simultaneously that's why it exceeds the

3:02

rate limit but thanks to another

3:03

YouTuber called Anubis AI he actually

3:06

modified the main.p code and add some

3:09

delay between every API call so it won't

3:11

trigger this Arrow I'll put a modified

3:13

code in the description below so that

3:15

you can copy and paste in here just

3:17

override the whole main.py file now we

2 Create product requirements

3:20

finish the setup the next step is get

3:22

the product requirement you can write

3:24

simple product requirement by yourself

3:25

but to make sure small AI actually

3:28

produce the results you want you

3:30

actually need to add a lot more details

3:31

to it what I normally do is I will go to

3:33

chatgpt and ask gpd4 to help me write

3:37

down the whole product requirement doc

3:39

which including the core functionalities

3:41

interface descriptions as well as the

3:44

file structure and I can just copy those

3:46

requirements put into this MD file that

3:48

we just created and then run this model

3:50

around ming.py prompt prd.md you can see

3:55

here it already identified what kind of

3:56

files they need to create as well as

3:59

break down all the shared dependencies

4:01

including the variables data schema Dom

4:04

element and function names this whole

4:07

process is going to take a while for

4:08

some complicated apps it can take like

4:10

15 to even 40 minutes but here is also

4:13

another tips I want to share what I

4:15

noticed that at this stage is still

4:17

pretty buggy in terms of making sure the

4:19

function writing in the file a is also

4:21

compatible with file B so for 90 of time

4:24

for the whole code base it generated it

4:25

didn't work out a box however if I asked

4:28

it to write everything into just one

4:30

file then it'll work magically like 80

4:33

to 90 of time the code I generate it

4:35

just work out of box so in the

4:37

requirement I will often say that it

4:38

should only have one file that include

4:40

everything and that's exactly what I did

4:42

here so now since it is finished the

3 Let Smol AI generate code

4:45

file is generated will be under this

4:47

generated folder if we run this it is a

4:51

snake game out of box it has all the

4:53

features that keep a snake move and it

4:55

has a function to generate food and also

4:58

have a way to keep a score and the last

4 Debug

5:00

thing I want to mention is that they not

5:02

only provides a way to generate a code

5:04

they actually provide a way to debug the

5:06

code as well so if I add another window

5:08

here I can run that command line called

5:10

Moto run debugger.py at the prompt

5:13

here's where I can add any Arrow I saw

5:15

when I'm running the code and the same

5:17

thing I can put the arrows here in

5:19

prompt directly but I can also create

5:21

debug.md and then you can add the arrows

5:24

you saw and we can run this model.round

5:26

debugger py prompt debug.md

5:30

thank you

5:31

and then you can see it will start

5:33

giving me analysis of the potential

5:35

arrows and how can I fix it so it's like

5:37

having both Engineers who generate code

5:39

and also a debugger at the same time

5:41

that's pretty cool and this is small AI

5:44

developer as I mentioned it is a little

5:46

bit buggy it's not capable to produce a

5:48

very complicated apps out of box yet so

5:51

I don't think it's going to replace a

5:52

real engineer with its current format

5:55

however as a product designer and

5:56

product manager I'm really excited about

5:58

this because if I want to user test a

6:01

certain product ideas I can ask it to

6:03

build a prototype very very quickly and

6:05

tasks with real users for almost zero

6:07

cost but on the other side you can think

6:09

about how this whole ecosystem start

6:12

composed on top of each other we already

6:14

see AI that can design the user

6:16

interface and it's very easy for AI to

6:19

produce product requirement Doc and do

6:21

research so I think it's very possible

6:23

that in future we can have the whole AI

6:25

product team that including both product

6:27

manager product designer engineer QA out

6:30

of box and care around lots of different

6:32

experiments for you autonomously so I'm

6:34

very excited to see what other

6:36

interesting apps that you asked small AI

6:38

to do and comment below for any ideas

6:39

you have thank you and see you next time

English (auto-generated)

All

From AI Jason

AI

Python

Related

For you

Listenable

Recently uploaded

Watched

0:00

as an engineer small developer is

0:02

absolutely amazing and simultaneously

0:05

terrifying it is like having a junior

0:08

engineer in your pocket this goes beyond

0:11

putting a prompt into chat gbt and

0:12

having it help you code this actually

0:14

creates an entire project for you and

0:17

it's super easy to use so I'm going to

0:19

show you a project that I created in two

0:21

minutes then I'm going to show you how

0:23

to install it let's do it so here's the

0:25

prompt that I gave it write a python

0:27

project that takes a Chachi PT API key

0:29

in an m-file and then when the main

0:32

script runs ask the user for a prompt

0:35

and then use the chat GPT 3.5 turbo API

0:38

endpoint to get a response from chat gbt

0:40

then it displays that response to the

0:43

user and asks for another prompt make

0:45

sure to include a requirements.txt file

0:47

also make sure to use the open AI module

0:50

here's an example of what the chat GPT

0:52

API call should look like and then I

0:54

gave it an example straight from the

0:56

Chachi PT API docs so I ran small

0:59

developer on this prompt and it output

1:01

an entire project for me so we have the

1:04

requirements file

1:06

we have the main file and we also have

1:09

an m-file now I'm definitely going to

1:10

rotate this API key before publishing

1:12

the video so let's run it let's see what

1:14

happens

1:16

enter your prompt tell me a joke there

1:19

it is why couldn't the bicycle stand up

1:21

by itself because it was too tired

1:23

amazing when was Bill Clinton President

1:25

Bill Clinton was the 42nd President of

1:28

the United States and he served two

1:29

terms from January 20th 1993 to January

1:32

20th 2001. so again I created this

1:35

entire project from a few lines of a

1:38

prompt and it actually created all the

1:40

files for me so let me show you how to

1:42

install it now so this is the GitHub

1:43

page small Dash AI slash developer it

1:46

has over 7 000 stars and nearly 500

1:49

forks and it's one of the trending

1:51

GitHub repos right now human-centric and

1:54

coherent whole program synthesis AKA

1:56

your own personal Junior developer so it

1:58

gives a bunch of information about what

2:00

it does but let's actually do it now

2:03

this is so easy to use and they give a

2:04

few examples of incredible projects

2:06

built from just a few prompts and you

2:08

can get extremely detailed in these

2:10

prompts you can think of it like writing

2:12

a spec and then you pass it off to a

2:14

junior developer to write and writes it

2:17

for you and creates all the files so

2:19

it's amazing so the first thing we're

2:21

going to do is come down around

2:22

two-thirds of the way through the page

2:23

we're going to grab this line git clone

2:25

and then the GitHub repo we're going to

2:27

copy that so I have a new vs code window

2:29

open I'm going to come up to the top

2:31

right and click for a new terminal now

2:33

once that new terminal opens I'm going

2:35

to CD to the desktop

2:37

hit enter and I'm going to paste that

2:39

line git clone github.com small Dash AI

2:42

slash developer then I'm going to hit

2:44

enter and then that's going to clone it

2:46

to my desktop from there I'm going to CD

2:49

into that folder so now I'm in the

2:51

folder so the next thing I'm going to do

2:53

is come up to the top left click the

2:54

Explorer icon and then click open folder

2:57

then I'm going to open the developer

2:58

folder so there it is that's the entire

3:00

small developer project I'm going to

3:02

open up the terminal Again by clicking

3:04

the toggle panel in the top right and

3:06

while that's going I'm going to rename

3:07

the DOT example.m

3:11

to just dot end

3:14

then we're going to click on it and

3:17

we're going to enter our openai API key

3:19

now you can use anthropic as well if you

3:21

want but I'm just going to stick with

3:23

openai right so if you don't have an

3:25

open AI API key head on over to open Ai

3:28

and just generate one and we're going to

3:30

save this file we're going to go down to

3:31

main.pi and this is the main file so the

3:34

nice thing about this project is it uses

3:35

modal.com I had actually not heard of

3:38

modal.com but essentially it's like a

3:40

container like Docker and it really

3:42

takes away all the complexities of

3:44

managing module versions python versions

3:46

which I always stumble on and according

3:49

to a lot of the comments in my videos a

3:51

lot of you stumble on too so this is a

3:53

great solution to that now you don't

3:54

need it but it really makes it easier so

3:56

to use modal go to modal.com sign up for

3:59

a new account it says here that it's in

4:01

private beta but I was able to sign up

4:03

no problem so I don't think it's in

4:04

private beta anymore then we're going to

4:06

copy this right here pip install modal

4:08

Dash client

4:10

I'm going to switch back to my terminal

4:11

and I'm going to paste it so pip install

4:14

modal Dash client enter I already have

4:17

it installed but if you didn't it would

4:19

have installed it there now if you

4:20

wanted to run this without modal all you

4:22

have to do is PIP install Dash R

4:24

requirements.txt and then run the file

4:27

python main no modal.pi rather than

4:31

main.pi but we're going to stick with

4:33

using modal and so the basic usage is

4:36

right here so we're going to grab just

4:38

these first commands and then I'm going

4:40

to copy it and it's modal run main.pi

4:43

dash dash prompt switch back to visual

4:46

studio code and I'm going to paste that

4:47

in and so let's start with something

4:49

really basic

4:51

write a ruby script that counts to 100.

4:57

and then I'm going to hit enter now the

4:59

first time you do this modal is going to

5:00

ask you to authenticate and all you have

5:02

to do is click the link open up the

5:04

website log in and then switch back to

5:07

terminal and you're done so it says it's

5:09

going to create one file count to 100.

5:11

RB and you can see here that it's

5:14

actually using containers with modal and

5:17

it has a pretty nice UI for being

5:19

strictly through the terminal and there

5:21

it is it actually outputs the file name

5:23

and what's in the file but the cool part

5:25

is it actually generates it for me so if

5:27

I come up here to the left and click

5:29

this generated folder click the drop

5:32

down there's the file it just created

5:34

count to 100. RB So Def count to 100 for

5:38

I and 1 to 100 puts I that is correct

5:42

now obviously this was a very simple

5:43

example but you can get quite complex

5:45

and the nice thing is you don't actually

5:48

have to put the entire prompt in the

5:50

terminal you can create this thing

5:51

called prompt.md and put your entire

5:54

prompt in there alright so I have my

5:56

little screen script here modal run

5:58

main.pi dash dash prompt and then I

6:01

output The Prompt so I'm going to

6:02

highlight that whole thing I'm going to

6:04

come to the terminal I'm going to paste

6:06

it and then when it's finished I hit

6:07

enter and now it's going to start

6:08

creating that project now you could just

6:10

create a DOT MD file and put your prompt

6:13

in there instead of having to paste the

6:15

entire prompt directly into terminal now

6:17

while that's going let me show you an

6:18

example of what they've done now here's

6:20

a really detailed spec of a Chrome

6:22

extension now I don't have access to

6:24

anthropic Claude yet but as soon as I do

6:27

I'm going to test this out but for now

6:29

you can see they basically created an

6:31

entire detailed spec that can be passed

6:33

to a junior developer and in this case

6:35

the junior developer is Chachi PT and

6:38

there it's done app completed so let's

6:40

check it out now we're going to go to

6:42

the generated folder we're going to

6:43

click it and we're going to look inside

6:45

so we have the project root

6:48

we have our dot m file we have the

6:51

main.pi file and it looks like

6:53

everything is correct and we even have

6:55

the requirements.txt file so for here we

6:59

need an API key so I'm going to go grab

7:01

that I'm going to double click it paste

7:03

and now I have the API key and I'm going

7:05

to save then I'm going to go to main.pi

7:07

and I'm going to push play and let's see

7:10

if it works okay so it loaded tell me a

7:12

joke hit enter and there it is why

7:15

wouldn't the bicycle stand up by itself

7:16

because it was too tired funny it gave

7:19

me the same exact joke as before and

7:21

that's it we've created an entire

7:22

project just with a prompt and it

7:24

creates the entire file structure all of

7:26

the code for it and it's easily done

7:30

give it a try let me know what you think

7:31

if you like this video please consider

7:33

giving me a like And subscribe and I'll

7:35

see you in the next one

English (auto-generated)

All

From Matthew Berman

AI

Computer programming

Related

For you

Listenable

Recently uploaded

Watched

0:00

foreign

0:10

so

0:13

[Music]

0:15

basically make sure you've got a GitHub

0:17

account uh make sure you have an IDE

0:19

installed so visual studio or pycharm

0:25

I'll use visual studio for this video

0:27

I'm typically a jetbrains user so might

0:31

have to struggle but most people use

0:33

Visual Studio

0:34

so you're going to go here to this uh

0:39

this URL right here

0:43

and go to code

0:46

copy that

0:47

[Music]

0:50

go ahead and create

0:52

a new folder if you run Windows just go

0:56

wherever you want to create

0:59

a folder

1:00

and let's call it small

1:05

[Music]

1:07

bit

1:09

okay go into that folder

1:11

click up here type in CMD it's going to

1:13

open a

1:15

command line interface right in the

1:17

folder that you need and then just type

1:19

git clone and then paste that URL that

1:22

we posted there earlier

1:25

hit enter

1:26

and it's going to create uh

1:29

new folder called developer

1:33

and you'll see that

1:36

your project has been installed there

1:37

okay so if you close out of that you see

1:39

this new developer folder

1:41

and here is our project okay

1:45

so go ahead and copy that open up your

1:48

IDE

1:50

a and open that folder that we just

1:54

copied

1:55

and this is the project

1:59

so first thing that you're going to want

2:01

to do is update your environment file

2:06

so we can duplicate this

2:08

all right actually just rename it it's

2:10

fine

2:11

and then you rename it just delete that

2:13

part and call it EnV so now it's active

2:16

you're gonna need an open AI API key and

2:19

I don't know if this is essential for

2:20

you to have but

2:23

I think you can get away of running this

2:26

without having an anthropic API key

2:29

so go ahead and slap in a couple API

2:33

keys there

2:34

by going to your open API account

2:38

create a new wand

2:48

that in there

2:53

and then go go to your anthropic account

2:55

it's going to use cloud

2:57

AI

2:59

and grab your API key if you got one

3:10

okay

3:12

put that in there

3:17

okay then save that file okay so the

3:19

next thing that we need to do is

3:22

go ahead and create a modal account

3:25

okay and when you create your modal

3:27

account sign up with your GitHub account

3:30

because it's going to make everything

3:31

hold hands and work together properly in

3:33

your project

3:34

okay

3:36

so do that so after you've signed up go

3:39

and click on this

3:41

and then you're gonna copy this

3:44

and you want to create a new token then

3:46

you're going to put that into your

3:50

what we're using here

3:54

uh terminal down here

3:58

okay

3:59

so paste that in there

4:06

and then what it's going to do is

4:08

connect your modal account

4:10

you see here create token

4:12

token created close this window

4:15

and now that's connected this

4:18

small AI developer to your to your

4:21

account okay so it's beautiful what it's

4:23

going to do is when it's creating files

4:25

it's going to put them up there and

4:26

create them there

4:27

and then at the end it'll it'll generate

4:30

everything you want here okay for code

4:33

okay

4:34

okay so we want to give the

4:37

small AI a good description of what it

4:40

is that we're trying to build so I'm

4:41

kind of just doing this on the Fly

4:43

trying to build a really simple CMS

4:45

client management system

4:47

um out of vanilla JavaScript HTML and

4:50

CSS okay so I've gone ahead and written

4:53

down what I want its functions to do

4:55

and uh you can get really elaborate with

4:58

this I built out extremely complicated

5:00

apps with this and there's another video

5:02

I have for how to make that happen

5:04

because it'll crash if you go beyond

5:06

this scale but I have code that'll allow

5:08

to run at that level uh in another video

5:12

so basically we want to describe in

5:15

detail what it is that we want to build

5:17

but instead of sticking this in the

5:19

prompt uh right away when we fire up

5:21

small AI what we're going to do is we're

5:24

going to come back and we're going to

5:26

change this uh default prompt here okay

5:29

let's just delete all this

5:31

and then we're going to have it when we

5:32

run it in the command line look at this

5:34

instead of the prompt that we put in the

5:36

command line and the power in this is

5:39

that it runs recursively so it'll go

5:41

back and uh

5:43

check that it's done these things and

5:45

added these features which is super cool

5:47

so we'll just go ahead and copy paste

5:49

that into our markdown file that's

5:53

called promptmd

5:54

okay and then hit save

5:57

and then next we're gonna run this or

6:00

sorry we're gonna run small

6:03

AI

6:04

developer

6:06

and how we're going to do that is we're

6:09

going to run this command

6:11

okay so make sure you're in the

6:13

developer folder

6:15

so modal run

6:18

mainpie.prompt and that's telling it to

6:19

look at this prompt markdown okay so

6:22

after you do that hit enter

6:24

and watch the magic happen

6:29

so

6:30

what's done is it's parsed in this

6:32

command okay

6:34

and then you can see if you click here

6:39

thank you

6:40

it's it's gonna start building this out

6:44

uh

6:45

to modal okay

6:48

so let's just watch everyone

6:52

all right so it's processing okay so

6:54

this is the magic I've worked so

6:56

it's identified like this is the

7:00

this is so powerful guys I can't even

7:01

get over it it's identified the

7:03

different uh separation of concerns that

7:05

we've got going on here okay and then

7:07

it's

7:09

identifying its dependencies is data

7:12

schema okay

7:15

and then it's creating all the elements

7:18

uh Dom IDs okay this this would take

7:21

hours and hours and hours to do all

7:24

right and then the different message

7:26

names

7:27

and now it's defining all the different

7:29

functions that it's going to need

7:31

okay and then it's updating all this

7:33

stuff

7:34

boom boom boom boom here it is

7:36

processing

7:37

back and forth it does this

7:39

asynchronously in parallel okay now it's

7:42

generating all the functions

7:44

in these different javascripts boom boom

7:47

boom boom go

7:49

uh it's still running it's building it

7:51

but you wait and see this is really

7:54

amazing stuff

7:56

okay so the app is complete that took

7:58

like not even three minutes

8:00

okay

8:03

this is

8:05

wow wow wow

8:07

okay so here's rcss

8:10

that generated and we're not talking

8:12

like Mickey Mouse like it's done the

8:14

whole thing as a hundred lines

8:16

of CSS okay here's our JavaScript

8:19

there's our app

8:20

there's our

8:22

JavaScript for uh managing the clients

8:26

interactions and whatnots

8:28

okay and then here's our index okay

8:31

so that you can interact with this

What is Artificial Intelligence and how does it work? What are the 3 types of AI? The 3 types of AI are: General AI: AI that can perform all of the intellectual tasks a human can. Currently, no form of AI can think abstractly or develop creative ideas in the same ways as humans.  Narrow AI: Narrow AI commonly includes visual recognition and natural language processing (NLP) technologies. It is a powerful tool for completing routine jobs based on common knowledge, such as playing music on demand via a voice-enabled device.  Broad AI: Broad AI typically relies on exclusive data sets associated with the business in question. It is generally considered the most useful AI category for a business. Business leaders will integrate a broad AI solution with a specific business process where enterprise-specific knowledge is required.  How can artificial intelligence be used in business? AI is providing new ways for humans to engage with machines, transitioning personnel from pure digital experiences to human-like natural interactions. This is called cognitive engagement.  AI is augmenting and improving how humans absorb and process information, often in real-time. This is called cognitive insights and knowledge management. Beyond process automation, AI is facilitating knowledge-intensive business decisions, mimicking complex human intelligence. This is called cognitive automation.  What are the different artificial intelligence technologies in business? Machine learning, deep learning, robotics, computer vision, cognitive computing, artificial general intelligence, natural language processing, and knowledge reasoning are some of the most common business applications of AI.  What is the difference between artificial intelligence and machine learning and deep learning? Artificial intelligence (AI) applies advanced analysis and logic-based techniques, including machine learning, to interpret events, support and automate decisions, and take actions.  Machine learning is an application of artificial intelligence (AI) that provides systems the ability to automatically learn and improve from experience without being explicitly programmed.  Deep learning is a subset of machine learning in artificial intelligence (AI) that has networks capable of learning unsupervised from data that is unstructured or unlabeled.  What are the current and future capabilities of artificial intelligence? Current capabilities of AI include examples such as personal assistants (Siri, Alexa, Google Home), smart cars (Tesla), behavioral adaptation to improve the emotional intelligence of customer support representatives, using machine learning and predictive algorithms to improve the customer’s experience, transactional AI like that of Amazon, personalized content recommendations (Netflix), voice control, and learning thermostats.  Future capabilities of AI might probably include fully autonomous cars, precision farming, future air traffic controllers, future classrooms with ambient informatics, urban systems, smart cities and so on.  To know more about the scope of artificial intelligence in your business, please connect with our expert.

מהי בינה מלאכותית וכיצד היא פועלת? מהם 3 סוגי הבינה המלאכותית? שלושת סוגי הבינה המלאכותית הם: בינה מלאכותית כללית: בינה מלאכותית שיכולה לבצע את כל המשימות האינטלקטואליות שאדם יכול. נכון לעכשיו, שום צורה של AI לא יכולה לחשוב בצורה מופשטת או לפתח רעיונות יצירתיים באותן דרכים כמו בני אדם. בינה מלאכותית צרה: בינה מלאכותית צרה כוללת בדרך כלל טכנולוגיות זיהוי חזותי ועיבוד שפה טבעית (NLP). זהו כלי רב עוצמה להשלמת עבודות שגרתיות המבוססות על ידע נפוץ, כגון השמעת מוזיקה לפי דרישה באמצעות מכשיר התומך בקול. בינה מלאכותית רחבה: בינה מלאכותית רחבה מסתמכת בדרך כלל על מערכי נתונים בלעדיים הקשורים לעסק המדובר. זה נחשב בדרך כלל לקטגוריית הבינה המלאכותית השימושית ביותר עבור עסק. מנהיגים עסקיים ישלבו פתרון AI רחב עם תהליך עסקי ספציפי שבו נדרש ידע ספציפי לארגון. כיצד ניתן להשתמש בבינה מלאכותית בעסק? AI מספקת דרכים חדשות לבני אדם לעסוק במכונות, ומעבירה את הצוות מחוויות דיגיטליות טהורות לאינטראקציות טבעיות דמויות אדם. זה נקרא מעורבות קוגניטיבית. בינה מלאכותית מגדילה ומשפרת את האופן שבו בני אדם קולטים ומעבדים מידע, לעתים קרובות בזמן אמת. זה נקרא תובנות קוגניטיביות וניהול ידע. מעבר לאוטומציה של תהליכים, AI מאפשר החלטות עסקיות עתירות ידע, תוך חיקוי אינטליגנציה אנושית מורכבת. זה נקרא אוטומציה קוגניטיבית. מהן טכנולוגיות הבינה המלאכותית השונות בעסק? למידת מכונה, למידה עמוקה, רובוטיקה, ראייה ממוחשבת, מחשוב קוגניטיבי, בינה כללית מלאכותית, עיבוד שפה טבעית וחשיבת ידע הם חלק מהיישומים העסקיים הנפוצים ביותר של AI. מה ההבדל בין בינה מלאכותית ולמידת מכונה ולמידה עמוקה? בינה מלאכותית (AI) מיישמת ניתוח מתקדמות וטכניקות מבוססות לוגיקה, כולל למידת מכונה, כדי לפרש אירועים, לתמוך ולהפוך החלטות לאוטומטיות ולנקוט פעולות. למידת מכונה היא יישום של בינה מלאכותית (AI) המספק למערכות את היכולת ללמוד ולהשתפר מניסיון באופן אוטומטי מבלי להיות מתוכנתים במפורש. למידה עמוקה היא תת-קבוצה של למידת מכונה בבינה מלאכותית (AI) שיש לה רשתות המסוגלות ללמוד ללא פיקוח מנתונים שאינם מובנים או ללא תווית. מהן היכולות הנוכחיות והעתידיות של בינה מלאכותית? היכולות הנוכחיות של AI כוללות דוגמאות כמו עוזרים אישיים (Siri, Alexa, Google Home), מכוניות חכמות (Tesla), התאמה התנהגותית לשיפור האינטליגנציה הרגשית של נציגי תמיכת לקוחות, שימוש בלמידת מכונה ואלגוריתמים חזויים כדי לשפר את חווית הלקוח, עסקאות בינה מלאכותית כמו זו של אמזון, המלצות תוכן מותאמות אישית (Netflix), שליטה קולית ותרמוסטטים ללמידה. יכולות עתידיות של AI עשויות לכלול כנראה מכוניות אוטונומיות מלאות, חקלאות מדויקת, בקרי תעבורה אוויריים עתידיים, כיתות עתידיות עם אינפורמטיקה סביבתית, מערכות עירוניות, ערים חכמות וכן הלאה. כדי לדעת יותר על היקף הבינה המלאכותית בעסק שלך, אנא צור קשר עם המומחה שלנו.

Google Colab: Google Colab היא פלטפורמה מקוונת המאפשרת למשתמשים לשתף ולהריץ סקריפטים של Python בענן Graphics Processing Unit(GPU): GPU, או יחידת עיבוד גרפית, הוא סוג מיוחד של שבב מחשב שנועד להתמודד עם המורכבות חישובים הדרושים להצגת תמונות ווידאו במחשב או במכשיר אחר. זה כמו המוח של המערכת הגרפית של המחשב שלך, והוא ממש טוב לעשות הרבה מתמטיקה ממש מהר. GPUs משמשים סוגים רבים ושונים של מכשירים, כולל מחשבים, טלפונים וקונסולות משחקים. הם שימושיים במיוחד למשימות הדורשות כוח עיבוד רב, כמו משחקי וידאו, עיבוד גרפיקה תלת-ממדית או הפעלת אלגוריתמים של למידת מכונה. מודל שפה גדול (LLM): סוג של מודל למידת מכונה שאומן על כמות גדולה מאוד של נתוני טקסט ומסוגל ליצור טקסט בעל צליל טבעי. Machine Learning (ML): שיטה ללמד מחשבים ללמוד מנתונים, מבלי להיות מתוכנתים במפורש. עיבוד שפה טבעית (NLP): תת-תחום של AI המתמקד בהוראת מכונות להבין, לעבד וליצור שפה אנושית רשתות עצביות: סוג של אלגוריתם למידת מכונה המבוססת על המבנה והתפקוד של המוח. שדות קרינה עצביים (NeRF): שדות קרינה עצביים הם סוג של מודל למידה עמוקה שיכול לשמש למגוון משימות, כולל יצירת תמונה, זיהוי אובייקטים ופילוח. NeRFs שואבים השראה מהרעיון של שימוש ברשת עצבית למודל של זוהר תמונה, שהוא מדד לכמות האור שנפלט או מוחזר על ידי אובייקט. OpenAI: OpenAI הוא מכון מחקר המתמקד בפיתוח וקידום טכנולוגיות בינה מלאכותית שהן בטוחות, שקופות ומועילות לחברה. Overfitting: בעיה נפוצה בלמידת מכונה, שבה המודל מתפקד היטב בנתוני האימון אך גרועים בחדשים, בלתי נראים. נתונים. זה מתרחש כאשר המודל מורכב מדי ולמד יותר מדי פרטים מנתוני האימון, כך שהוא לא מכליל היטב. הנחיה: הנחיה היא פיסת טקסט המשמשת לתכנון מודל שפה גדול ולהנחות את הדור שלו Python: Python היא שפת תכנות פופולרית ברמה גבוהה הידועה בפשטות, בקריאות ובגמישות שלה (כלי AI רבים משתמשים בה) למידת חיזוק: סוג של למידת מכונה שבה המודל לומד על ידי ניסוי וטעייה, מקבל תגמולים או עונשים על מעשיו ומתאים את התנהגותו בהתאם. מחשוב מרחבי: מחשוב מרחבי הוא השימוש בטכנולוגיה כדי להוסיף מידע וחוויות דיגיטליות לעולם הפיזי. זה יכול לכלול דברים כמו מציאות רבודה, שבה מידע דיגיטלי מתווסף למה שאתה רואה בעולם האמיתי, או מציאות מדומה, שבה אתה יכול לשקוע במלואו בסביבה דיגיטלית. יש לו שימושים רבים ושונים, כמו בחינוך, בידור ועיצוב, והוא יכול לשנות את האופן שבו אנו מתקשרים עם העולם ואחד עם השני. דיפוזיה יציבה: דיפוזיה יציבה מייצרת תמונות אמנותיות מורכבות המבוססות על הנחיות טקסט. זהו מודל AI של סינתזת תמונות בקוד פתוח הזמין לכולם. ניתן להתקין את ה-Stable Diffusion באופן מקומי באמצעות קוד שנמצא ב-GitHub או שישנם מספר ממשקי משתמש מקוונים הממנפים גם מודלים של Stable Diffusion. למידה מפוקחת: סוג של למידת מכונה שבה נתוני האימון מסומנים והמודל מאומן לבצע תחזיות על סמך היחסים בין נתוני הקלט והתוויות המתאימות. למידה ללא פיקוח: סוג של למידת מכונה שבה נתוני האימון אינם מסומנים, והמודל מאומן למצוא דפוסים ויחסים בנתונים בעצמו. Webhook: Webhook הוא דרך של תוכנת מחשב אחת לשלוח הודעה או נתונים לתוכנית אחרת דרך האינטרנט בזמן אמת. זה עובד על ידי שליחת ההודעה או הנתונים לכתובת URL ספציפית, השייכת לתוכנית האחרת. Webhooks משמשים לעתים קרובות כדי להפוך תהליכים לאוטומטיים ולהקל על תוכניות שונות לתקשר ולעבוד יחד. הם כלי שימושי למפתחים שרוצים לבנות יישומים מותאמים אישית או ליצור אינטגרציות בין מערכות תוכנה שונות.