1
00:00:00,150 --> 00:00:03,570
-: This lesson is all about GPT Vision

2
00:00:03,570 --> 00:00:07,200
and the ability to upload pictures into ChatGPT

3
00:00:07,200 --> 00:00:11,130
and have it give you real world useful feedback

4
00:00:11,130 --> 00:00:12,420
on those images.

5
00:00:12,420 --> 00:00:14,430
So let's dive right in.

6
00:00:14,430 --> 00:00:18,600
Here, I made an image in DALLE-3 with the help of ChatGPT,

7
00:00:18,600 --> 00:00:20,280
and it has a bunch of floating candles.

8
00:00:20,280 --> 00:00:24,116
So for starters, you just click here, attach image,

9
00:00:24,116 --> 00:00:26,790
and you just select an image you wanna upload and hit open,

10
00:00:26,790 --> 00:00:28,950
and then you can add a prompt,

11
00:00:28,950 --> 00:00:30,870
such as could say, what am I looking at

12
00:00:30,870 --> 00:00:32,880
or you could say, how do I fix this code

13
00:00:32,880 --> 00:00:35,340
because this is an image of VS Code.

14
00:00:35,340 --> 00:00:38,736
Once you have your image and your prompt in,

15
00:00:38,736 --> 00:00:41,520
you just hit enter and watch ChatGPT do all the rest.

16
00:00:41,520 --> 00:00:44,130
Here we have this image and I said,

17
00:00:44,130 --> 00:00:46,680
how many floating candles are in this photo?

18
00:00:46,680 --> 00:00:48,750
It says there are approximately 60.

19
00:00:48,750 --> 00:00:51,420
Of course it says it's a little bit hard to tell,

20
00:00:51,420 --> 00:00:54,480
but it is clearly able to see what we're looking at here.

21
00:00:54,480 --> 00:00:56,040
I drew a picture,

22
00:00:56,040 --> 00:00:59,550
and this is supposed to be the general outline of a website,

23
00:00:59,550 --> 00:01:01,440
and I have this prompt here

24
00:01:01,440 --> 00:01:02,670
saying you're an expert web developer.

25
00:01:02,670 --> 00:01:06,660
Please write me the code in Next.js plus Tailwind CSS

26
00:01:06,660 --> 00:01:08,610
and I said, take a deep breath

27
00:01:08,610 --> 00:01:11,370
and think step by step about how you'll do this.

28
00:01:11,370 --> 00:01:12,960
Now write the complete code

29
00:01:12,960 --> 00:01:15,510
for this working one step at a time.

30
00:01:15,510 --> 00:01:18,060
And just a quick little side note here,

31
00:01:18,060 --> 00:01:19,620
when you say take a deep breath,

32
00:01:19,620 --> 00:01:22,470
think step by step about how you will do this,

33
00:01:22,470 --> 00:01:24,420
ChatGPT has been showing over time

34
00:01:24,420 --> 00:01:26,700
that it is really good at prompting itself.

35
00:01:26,700 --> 00:01:28,620
So when you give it the instruction

36
00:01:28,620 --> 00:01:32,100
and allow it to generate some of its own instruction,

37
00:01:32,100 --> 00:01:34,980
it will only make the result better.

38
00:01:34,980 --> 00:01:37,620
Now we have a breakdown of

39
00:01:37,620 --> 00:01:41,008
what we're seeing here on the piece of paper

40
00:01:41,008 --> 00:01:42,300
and a step-by-step plan.

41
00:01:42,300 --> 00:01:45,960
It then helps us set up a new project

42
00:01:45,960 --> 00:01:50,820
and give us the code that we need to create the website

43
00:01:50,820 --> 00:01:52,860
that I asked it to create.

44
00:01:52,860 --> 00:01:56,520
Now, what's even more insane to me

45
00:01:56,520 --> 00:01:58,770
is that not only did it give us the code,

46
00:01:58,770 --> 00:02:02,400
but I haven't used those specific coding languages before

47
00:02:02,400 --> 00:02:06,360
and it was walking me step by step what I needed to do

48
00:02:06,360 --> 00:02:09,539
in order to make this happen.

49
00:02:09,539 --> 00:02:13,980
So I was uploading pictures of VS Code

50
00:02:13,980 --> 00:02:15,900
and asking ChatGPT

51
00:02:15,900 --> 00:02:20,670
where in this interface do I paste the code that it gave me?

52
00:02:20,670 --> 00:02:22,620
We went through the step-by-step process

53
00:02:22,620 --> 00:02:24,992
of getting me set up properly

54
00:02:24,992 --> 00:02:26,460
in order to be able to run this,

55
00:02:26,460 --> 00:02:29,250
getting myself familiarized with the screen here,

56
00:02:29,250 --> 00:02:32,700
and then actually started coding together

57
00:02:32,700 --> 00:02:36,903
in a coding language that before this I have never used.

58
00:02:37,779 --> 00:02:41,460
So it told me exactly where to paste the code,

59
00:02:41,460 --> 00:02:43,590
which parts needed to change.

60
00:02:43,590 --> 00:02:45,480
We did some troubleshooting together.

61
00:02:45,480 --> 00:02:48,813
I would upload an image and say, hey, there's an error.

62
00:02:49,664 --> 00:02:51,792
What went wrong, what do I do?

63
00:02:51,792 --> 00:02:55,530
And it was able to get a working finished product

64
00:02:55,530 --> 00:02:58,680
from here by me just uploading images

65
00:02:58,680 --> 00:03:02,640
of my screen, step-by-step every time I got stuck,

66
00:03:02,640 --> 00:03:04,950
which is really incredible

67
00:03:04,950 --> 00:03:07,680
because now there is no barrier to entry

68
00:03:07,680 --> 00:03:10,600
to learn literally anything.

69
00:03:10,600 --> 00:03:12,450
Any software, anything that's on your computer,

70
00:03:12,450 --> 00:03:14,490
even if you've never seen it before,

71
00:03:14,490 --> 00:03:16,830
you can just upload a picture to ChatGPT

72
00:03:16,830 --> 00:03:18,690
and say, hey, how do I begin?

73
00:03:18,690 --> 00:03:20,820
Where do I start or I'm stuck with this

74
00:03:20,820 --> 00:03:22,650
or just take a picture of the problem

75
00:03:22,650 --> 00:03:25,320
and it will probably intelligently understand

76
00:03:25,320 --> 00:03:26,580
what you're trying to do,

77
00:03:26,580 --> 00:03:30,510
which is really just so, so incredible.

78
00:03:30,510 --> 00:03:32,520
I think this feature alone

79
00:03:32,520 --> 00:03:37,520
is one of the most mind blowing things that I've seen AI do,

80
00:03:37,530 --> 00:03:39,990
and we've all seen all of these amazing features

81
00:03:39,990 --> 00:03:40,863
come to life,

82
00:03:42,136 --> 00:03:44,792
but now if I can code something in a language

83
00:03:44,792 --> 00:03:48,150
that I haven't used before, the sky is the limit.

84
00:03:48,150 --> 00:03:51,510
Simply upload whatever you need to to ChatGPT

85
00:03:51,510 --> 00:03:55,170
and have it help you fix it, have it help you organize it.

86
00:03:55,170 --> 00:03:58,184
You could take a picture of ingredients and say,

87
00:03:58,184 --> 00:03:59,712
what can I cook with this?

88
00:03:59,712 --> 00:04:01,240
You could take a picture of food

89
00:04:01,240 --> 00:04:02,100
and say how did somebody make this?

90
00:04:02,100 --> 00:04:03,750
The sky is the limit

91
00:04:03,750 --> 00:04:06,240
and I hope you really just start playing around

92
00:04:06,240 --> 00:04:09,840
and see all the amazing use cases that you can use

93
00:04:09,840 --> 00:04:11,463
with this new feature.

