The Sad Reality of AI Job Market w/ ML Engineer

Goda Go
15 Jul 202413:53

Summary

TLDRDans ce script de vidéo, un ingénieur en apprentissage automatique partage ses expériences et les prévisions du marché de l'emploi dans l'IA. Il discute des malentendus courants autour de l'IA, la difficulté de l'intégration des modèles ML dans les infrastructures des entreprises et les compétences manquantes dans les formations académiques. Il met en lumière l'importance de l'ingénierie de l'intégration et de la maintenance des modèles, ainsi que l'émergence des compétences en ingénierie de prompt pour les modèles génératifs d'IA.

Takeaways

  • 🎓 L'auteur a choisi de suivre une maîtrise pendant l'essor de l'IA, sachant que le domaine était en pleine croissance.
  • 📈 Il y a eu une forte demande pour les ingénieurs en apprentissage automatique et la science des données, mais celle-ci a diminué en 2023, suite à une surévaluation de l'IA en 2022.
  • 🤖 La communauté a une tendance à surestimer les capacités de l'IA comme étant une solution miracle, ce qui n'est pas le cas.
  • 🏢 Les entreprises ont souvent besoin d'ingénieurs qui peuvent non seulement créer des modèles de machine learning, mais aussi les intégrer efficacement dans leur infrastructure existante.
  • 💼 Les salaires des ingénieurs en apprentissage automatique sont élevés, et leur productivité doit justifier ces coûts.
  • 🔍 L'intégration de l'IA dans les entreprises est un défi majeur, nécessitant des compétences d'ingénierie pour gérer les pipelines de données et l'hébergement des modèles.
  • 👨‍🏫 L'éducation traditionnelle en apprentissage automatique se concentre plus sur la théorie que sur les compétences pratiques nécessaires pour l'industrie.
  • 📉 En 2023, les postes de machine learning ont diminué, tandis que ceux liés à l'intelligence artificielle générative et à l'ingénierie de prompts ont augmenté.
  • 🛠️ L'ingénierie de prompts est une compétence clé pour tirer le meilleur parti des modèles de l'IA, et elle n'est pas souvent enseignée dans les écoles.
  • 🔄 La maintenance des modèles de machine learning est importante, car leur performance peut dériver avec le temps si elle n'est pas surveillée et retraînée.
  • 🧐 L'auteur a découvert par lui-même l'importance de l'ingénierie de prompts en utilisant des modèles de l'IA pour résoudre des problèmes techniques complexes.

Q & A

  • Quelles étaient les attentes initiales de Tanner concernant l'industrie de l'IA lorsqu'il a commencé son master?

    -Tanner pensait qu'il terminerait son master au moment parfait, avec une forte demande pour des ingénieurs en apprentissage automatique, ce qui lui permettrait de décrocher un emploi facilement, de gagner beaucoup d'argent et de prendre sa retraite dans dix ans.

  • Quels sont les malentendus courants sur l'IA que Tanner a rencontrés dans son travail?

    -Un malentendu courant est que l'IA est une solution miracle qui peut augmenter rapidement les revenus de l'entreprise. En réalité, il est beaucoup plus difficile de créer une valeur durable avec l'IA qu'on ne le pense.

  • Pourquoi Tanner pense-t-il que les entreprises ont embauché massivement en 2022?

    -Les entreprises ne voulaient pas être laissées pour compte dans la course à l'IA, elles ont donc embauché massivement sans nécessairement savoir comment utiliser ces talents, par crainte que leurs concurrents ne les devancent.

  • Quelles compétences sont souvent négligées dans les programmes universitaires en IA, selon Tanner?

    -Les programmes universitaires se concentrent souvent sur la théorie et les mathématiques avancées nécessaires pour la recherche, mais négligent les compétences pratiques nécessaires pour intégrer les modèles dans les infrastructures complexes des entreprises.

  • Quel est l'impact de la formation axée sur la théorie par rapport à la formation axée sur la pratique, selon Tanner?

    -Une formation axée sur la théorie prépare mieux les étudiants à poursuivre des études académiques ou de recherche, mais ne les prépare pas suffisamment à appliquer leurs compétences de manière pratique dans l'industrie.

  • Comment Tanner a-t-il découvert l'ingénierie de prompts (prompt engineering)?

    -Tanner a découvert l'ingénierie de prompts en essayant de comprendre un bloc de code complexe avec l'aide de chatGPT. Cela a changé sa perception de l'utilité de chatGPT et l'a poussé à apprendre à mieux formuler ses requêtes.

  • Pourquoi Tanner pense-t-il que l'intégration des modèles d'IA est plus compliquée que leur création?

    -L'intégration des modèles d'IA est complexe car elle nécessite de l'ingénierie pour héberger le modèle, gérer les pipelines de données, assurer des prédictions rapides et surveiller les performances du modèle au fil du temps.

  • Quelles sont les nouvelles tendances en matière d'emplois en IA en 2023, selon Tanner?

    -En 2023, les emplois liés à l'IA générative et à l'ingénierie de prompts ont augmenté, tandis que la demande pour les ingénieurs en apprentissage automatique a diminué. Les compétences en chatGPT et l'ingénierie de prompts sont particulièrement recherchées.

  • Comment Tanner gère-t-il les attentes irréalistes des entreprises concernant l'IA?

    -Tanner explique aux entreprises que la création de valeur avec l'IA est plus difficile qu'elles ne le pensent et nécessite des compétences pratiques en ingénierie, pas seulement en modélisation.

  • Quel conseil Tanner donne-t-il à ceux qui veulent améliorer leurs compétences en ingénierie de prompts?

    -Tanner conseille d'expérimenter avec chatGPT pour voir ce qu'il peut faire, d'apprendre des techniques avancées d'ingénierie de prompts et de s'inspirer du travail de prompters expérimentés.

Outlines

00:00

🤖 L'engouement pour l'IA et ses déceptions

Dans le premier paragraphe, l'interviewé partage son expérience personnelle en tant qu'ingénieur en apprentissage automatique et scientifique des données. Il parle de l'engouement actuel pour l'IA et comment il a anticipé une demande forte pour les ingénieurs en apprentissage automatique après son diplôme. Cependant, la réalité s'est avérée différente, avec une demande accrue pour des postes exigeant une expérience avancée et des publications. Il souligne également que la demande a diminué en 2023, contrairement à l'engouement de 2022, et que les entreprises ont souvent des attentes irréalistes concernant l'IA, considérant souvent cette technologie comme une solution miracle. Il met en lumière les défis de l'intégration de l'IA dans les entreprises et la nécessité de compétences pratiques pour déployer l'IA de manière durable et efficace.

05:02

🔌 L'intégration de l'IA : un défi majeur

Le deuxième paragraphe se concentre sur les défis de l'intégration de l'IA dans les infrastructures des entreprises. L'interviewé explique que l'IA n'est pas une solution universelle et que son intégration nécessite des compétences d'ingénierie. Il aborde les questions de l'hébergement des modèles, la gestion des pipelines de données, la nécessité de nettoyer et de stocker les données, et la performance des modèles. Il souligne également l'importance de la surveillance continue des modèles et de la planification de leur réentraînement. L'interviewé critique l'éducation traditionnelle qui ne prépare pas les étudiants aux réalités de l'industrie et met en avant l'évolution des rôles dans le domaine de l'IA, notamment l'augmentation des postes liés à l'intelligence artificielle générative et à l'ingénierie de prompts.

10:02

💡 L'ingénierie de prompts : une compétence émergente

Dans le troisième paragraphe, l'interviewé partage son parcours personnel dans l'ingénierie de prompts, une compétence qui est devenue cruciale dans l'utilisation des modèles d'IA. Il raconte comment il a découvert par hasard les capacités de chatGPT pour aider à comprendre le code et comment cela a changé sa perception de l'IA. Il décrit son apprentissage des techniques d'ingénierie de prompts et comment cela a augmenté sa productivité et sa compréhension des modèles d'IA. L'interviewé insiste sur l'importance de l'apprentissage continu et de l'adaptation aux nouvelles compétences requises dans le domaine de l'IA, soulignant que l'ingénierie de prompts est une compétence universelle qui peut être appliquée à de nombreuses tâches professionnelles.

Mindmap

Keywords

💡Intelligence Artificielle (IA)

L'Intelligence Artificielle (IA) est un domaine de l'informatique qui vise à créer des machines capables de réaliser des tâches qui nécessitent normalement l'intelligence humaine, telles que la reconnaissance de la parole, la vision par ordinateur et la résolution de problèmes. Dans le script, l'IA est abordée comme un sujet de préoccupation majeur pour les entreprises qui cherchent à innover et à se démarquer sur le marché, mais aussi comme une source de malentendus et de défis dans la mise en œuvre réelle.

💡Machine Learning

Le Machine Learning est une branche de l'IA qui utilise des algorithmes pour permettre aux ordinateurs d'apprendre à partir de données sans être explicitement programmés. Dans le script, il est mentionné comme une compétence en demande pour les ingénieurs qui travaillent sur des projets d'IA, mais aussi comme un domaine où les attentes peuvent être déçues en raison de la difficulté de mettre en œuvre des modèles dans des environnements réels.

💡Data Scientist

Un Data Scientist est un professionnel qui analyse des données volumineuses pour extraire des informations utiles et prendre des décisions éclairées. Dans le script, le rôle du Data Scientist est associé à celui de l'ingénieur en apprentissage automatique, soulignant l'importance de la gestion des données et de l'analyse statistique dans le développement d'applications IA.

💡Attentes

Les attentes dans le script se réfèrent aux anticipations et aux espérances des employeurs et des employés concernant les capacités et les performances de l'IA. Il est souligné que ces attentes peuvent être déçues en raison de la complexité de l'intégration de l'IA dans les entreprises et des défis associés à la gestion des modèles d'apprentissage automatique.

💡Intégration

L'intégration dans le script fait référence au processus de mise en place des modèles d'apprentissage automatique dans les infrastructures existantes des entreprises. Il est souligné que cette étape est souvent plus complexe que prévu et nécessite des compétences techniques spécifiques pour s'assurer que les modèles fonctionnent correctement et apportent une valeur ajoutée.

💡Modèles d'Apprentissage Automatique

Les modèles d'apprentissage automatique sont des algorithmes utilisés pour analyser des données et faire des prédictions. Dans le script, ils sont mentionnés comme un élément clé dans le développement de solutions IA, mais aussi comme une source de défis en termes de maintenance et d'intégration dans les systèmes informatiques des entreprises.

💡Infrastructure

L'infrastructure dans le script se réfère aux systèmes et aux processus techniques qui permettent de supporter les applications IA. Il est souligné que l'hébergement, la gestion des données et la maintenance des modèles sont des aspects cruciaux qui doivent être pris en compte pour garantir la performance et la stabilité des solutions IA.

💡Maintenance

La maintenance dans le script fait référence à la gestion continue des modèles d'apprentissage automatique pour s'assurer qu'ils restent performants et pertinents. Il est souligné que cela nécessite une surveillance régulière et des mises à jour fréquentes pour adapter les modèles aux changements dans les données et les exigences des utilisateurs.

💡Éducation Académique

L'éducation académique est mentionnée dans le script comme un moyen d'acquérir les connaissances théoriques nécessaires pour travailler dans le domaine de l'IA. Cependant, il est souligné que cette formation peut ne pas être suffisamment pratique pour préparer les étudiants à la réalité du travail dans les entreprises, où des compétences plus techniques et de gestion sont souvent requises.

💡Prompt Engineering

Le Prompt Engineering est une technique utilisée pour optimiser les interactions avec des systèmes d'IA, en particulier les assistants virtuels comme chatGPT. Dans le script, il est présenté comme une compétence en croissance et essentielle pour tirer le meilleur parti des modèles IA, en apprenant à formuler des questions et des instructions de manière à obtenir des réponses précises et utiles.

Highlights

The speaker anticipated a booming job market in AI upon graduation, expecting high demand and lucrative salaries, but reality did not meet those expectations.

AI and machine learning hype in 2022 led to inflated expectations, with many companies rushing to hire in the field, only to face a dip in demand in 2023.

Companies sought experienced AI professionals to build new teams, but the demand was not sustained, leading to a challenging job market for new graduates.

A common misconception is that AI is a 'silver bullet' for business growth, but integrating AI into a company's infrastructure is more complex than expected.

The speaker emphasizes the importance of not only building AI models but also the engineering skills required to integrate and maintain them within a company's ecosystem.

Educational institutions may not adequately prepare students with the practical skills needed for industry, focusing more on theory and academic research.

The speaker discusses the gap between academic training and the real-world application of AI, highlighting the need for a broader skill set.

AI job roles saw a decline in 2023, but generative AI and prompt engineering emerged as growing fields, with chatGPT becoming a valuable skill.

Prompt engineering is a widely applicable skill that can significantly increase productivity across various job roles.

The speaker's personal experience with prompt engineering began with using chatGPT to understand complex code, leading to a deeper exploration of its capabilities.

Learning to effectively use AI tools like chatGPT involves understanding how to craft prompts to extract the most value.

The speaker's experience with chatGPT revealed the importance of prompt engineering in controlling the output and avoiding 'hallucinations' in responses.

Experimenting with different prompt settings can lead to varied results, indicating the need for careful consideration in AI tool usage.

The speaker discusses the shift of AI professionals from academia to industry, driven by financial incentives and the allure of working with cutting-edge technology.

The importance of monitoring and maintaining AI models over time to ensure continued performance and relevance is highlighted.

The speaker reflects on the need for continuous learning and adaptation in the field of AI, as traditional education may not fully equip professionals for industry demands.

A course on advanced prompt engineering is mentioned, indicating a growing interest and need for specialized training in this area.

Transcripts

play00:00

I'm timing this perfectly, I'm gonna finish up

play00:02

in a year, or when I graduate, there'll just

play00:04

be 50 people begging to get the job, and I'll

play00:07

make a million dollars and retire in 10 years.

play00:10

And Tanner, hi, you're part of

play00:13

Simpline's team and you're a machine

play00:15

learning engineer and data scientist.

play00:17

I'm really curious first about the recent AI hype.

play00:21

In your experience, everything what you see

play00:24

as a machine learning engineer, what are

play00:26

common misconceptions people have about AI?

play00:29

Also, how do you manage those expectations when

play00:33

you have to work on a project or with a client?

play00:35

Yeah, so there's, I can either speak to my

play00:39

recent experience as someone trying to get a job

play00:41

in this space and then maybe we can translate

play00:43

to them more about what the job entails.

play00:45

So yeah, this did, it was very

play00:47

fortuitous timing for me to be a

play00:50

graduate around this time of AI hype.

play00:52

Machine learning was definitely

play00:53

growing in popularity when I made

play00:55

the transition to the master.

play00:57

That was a deliberate career move.

play01:00

And then when I was halfway through my

play01:02

degree, chatGPT came out, and all the image

play01:04

generators got really popular, and the hype

play01:07

was enormous, and I thought, great, I'm

play01:10

timing this perfectly, I'm gonna finish

play01:12

up in a year, or when I graduate, they'll

play01:14

just be 50 people begging to give me a job.

play01:17

And I'll make a million

play01:18

dollars and retire in 10 years.

play01:20

And that reality came

play01:22

crashing down pretty quickly.

play01:24

And I think the primary reason is that people,

play01:26

when I was looking for a job, every company

play01:29

wanted someone to either start their AI team and

play01:33

they wanted someone with 10 years of experience.

play01:35

Publications, PhD that sort of thing.

play01:39

So everyone wants machine learning engineering.

play01:41

Everyone wants AI, even if

play01:43

they might not know how.

play01:44

It's so in line with, uh,

play01:47

Stanford AI Index Report.

play01:48

The common misconception is that, oh

play01:50

my God, everyone is hiring in AI, but

play01:53

actually the boom was really in 2022.

play01:56

And then 2023 saw across the board,

play01:59

meaning it's not just U.S., but also other

play02:02

countries having like a dip in demand.

play02:04

From the report, they highlighted

play02:06

that big tech companies cut on

play02:08

their staff and we saw big layoffs.

play02:11

For me, the other one kind of obvious was that

play02:13

people probably already hired their machine

play02:16

learning engineers and build teams in 2022.

play02:20

And that demand was not as big as in 2023.

play02:24

What is your feeling why

play02:25

these job positions are down?

play02:28

And, yeah, obviously senior

play02:31

positions are in demand.

play02:32

I just think it's probably

play02:33

a combination of factors.

play02:35

I think you are correct in the sense that

play02:38

companies who maybe are a bit more rash

play02:41

with their decisions, didn't want to be

play02:43

left behind, didn't want to be left out.

play02:45

and just started hiring like crazy and probably

play02:47

said we'll figure out what to do with them later.

play02:49

But if we don't hire our competitors

play02:51

will, we don't want to be left behind.

play02:53

I think another aspect of it is that a lot

play02:56

of people think AI is like a silver bullet,

play03:00

where all you have to do is just hire a couple

play03:02

people like me and say "here's some data,

play03:05

go build something", and then you're going

play03:07

to increase your company revenue by 30%.

play03:10

And that is simply not the case.

play03:11

I think it's a lot.

play03:12

Harder than companies realize to

play03:14

actually build sustainable value for

play03:17

a company, especially a large company.

play03:20

And this is one of the problems in the

play03:21

industry and I think one of the problems

play03:23

with how machine learning is taught.

play03:25

I was taught a lot of theory about machine

play03:28

learning and AI, getting down to the

play03:29

nitty gritty of the math, really how it

play03:32

works, talking pretty high level linear

play03:34

algebra, calculus, probabilistics, and

play03:37

then theory on how these large models work.

play03:39

And that's all well and good for doing

play03:41

a research project for your degree.

play03:44

It's fine for academia, but in the real

play03:46

world, you need someone who can not only

play03:49

design a model and play around with the

play03:51

parameters until you get a good prediction

play03:53

on your test set, but actually integrate

play03:55

that into a generally pretty sophisticated

play03:58

and complicated company infrastructure.

play04:02

And that is the whole other set of skills that

play04:04

most people don't actually learn in school

play04:07

or really in other applications as well.

play04:11

Because people may go in with very solid knowledge

play04:13

and ability in terms of actually building models.

play04:16

They're much more ignorant about how to actually

play04:19

take a model that works and then integrate

play04:23

it into a company's software ecosystem.

play04:26

And then maintain that model to the point

play04:29

where it can actually produce substantial

play04:31

value for a company that it's worked for.

play04:34

Exorbitant price that these companies are paying.

play04:37

Salaries are quite high as well.

play04:39

The productivity of the modeling

play04:40

has to justify the salary.

play04:41

To us at synthminds.ai that it's in

play04:44

alignment what we saw in really 2023

play04:47

to be honest, that everyone wanted AI.

play04:50

But what we found that when you come

play04:51

in a company, they have all these

play04:53

existing softwares, all the processes.

play04:55

Not really even data strategy in that sense.

play04:58

And then they're like, Oh,

play04:59

just do this magical thing.

play05:01

And what becomes that, the

play05:04

integration becomes the challenge.

play05:07

I like what you bring up, uh, the integration,

play05:10

because that's, I don't think that's much

play05:12

talked about, because everyone is focused

play05:14

as just machine learning solves everything.

play05:17

Yeah.

play05:18

And the integration is way more complicated

play05:19

than people, I think, understand.

play05:21

And that's why there's actually a big

play05:22

push in the fields, in terms of company,

play05:26

they want people who are not just data

play05:27

scientists, who aren't just tinkerers with

play05:29

models, but people who are also engineers.

play05:31

And I think that's why it's a machine learning

play05:34

engineer, not just whipping up a model

play05:38

with PyTorch and getting good results on a

play05:41

predictive task and saying, okay, job well done.

play05:43

So we can create a model into an

play05:45

infrastructure, means you have to host this

play05:46

somewhere, same way you host a website.

play05:48

You have to figure out how the

play05:49

data pipeline is going to work.

play05:51

Are you ingesting data continuously,

play05:54

streaming it every now and then?

play05:55

How do you clean the data?

play05:57

How do you store the data?

play05:59

How do you then pass your data to the model?

play06:01

For a lot of companies, the time it takes

play06:04

to make a prediction critical may say,

play06:07

if you need a prediction fast, that is

play06:09

an engineering concern as well, because a

play06:10

lot of times these models are quite slow.

play06:13

So it's another engineering to figure out how

play06:15

are we going to handle millions of predictions

play06:16

at scale and still make sure the customer,

play06:18

the end user, is It's predicting quickly.

play06:22

How do you monitor a model?

play06:23

It changes over time and your model

play06:25

was trained on data from a year ago.

play06:27

Eventually your model performance, there's a

play06:28

good chance it drifts in a negative direction

play06:31

as in the performance gets worse over time.

play06:33

So how do you make sure you're

play06:35

monitoring the model's predictions

play06:36

and then do you have a plan to retrain the model?

play06:39

And these are all kinds of things that you don't

play06:41

learn usually at school, maybe at class, you

play06:44

don't learn it with other educational resources.

play06:47

They're arguably more important

play06:49

than being really good at creating a

play06:50

predictive model in the first place.

play06:52

Before machine learning engineers, would go

play06:55

into academia, like doing research, where

play06:58

all that theory can actually be applied and

play07:00

you can be tinkering with just math problems.

play07:02

But now, um, there is a, I think from 35 percent

play07:06

to 70 percent growth for engineers to actually

play07:10

go into industry, because this is where money

play07:13

is, this is where compute is, and it's much

play07:15

more interesting, but the traditional education,

play07:18

not necessarily prepared for this shift, right?

play07:22

I'm sure it's better at other schools

play07:23

and probably better at institutions that

play07:25

are more geared towards engineering,

play07:27

like engineering institutions.

play07:29

They probably teach you more of

play07:30

these practical skills virtually.

play07:32

I went to the McGill University, and then I also

play07:36

went to the research institute for my master's

play07:39

called Montreal Institute of Engineering.

play07:41

Learning Algorithm, called MILA.

play07:43

And both of these institutions

play07:44

care primarily about academia.

play07:46

They are trying to prepare their students

play07:48

to transition after this degree, to the

play07:50

next degree in their academic career.

play07:52

So, they want to prepare you to do a PhD.

play07:55

They don't necessarily want to prepare you to go

play07:57

take your skill and apply it at a large company.

play07:59

Which is why they focus so heavily on

play08:01

the theory, and things that are important

play08:03

for publishing scientific paper, and

play08:05

less so on the actual practical skills.

play08:07

And I imagine this is not just an

play08:08

issue in institutions I was at.

play08:10

Probably across the board in academia.

play08:12

As I mentioned to you that machine learning

play08:15

roles, they actually fell percentage wise in 2023.

play08:19

The roles that emerged and is actually growing

play08:22

is generative artificial intelligence, chatGPT,

play08:26

being a skill requirement on a job post.

play08:29

Prompt engineering closely followed that,

play08:32

and what was impressive to me that prompt

play08:34

engineering, there were more jobs in prompt

play08:36

engineering than in generative adversarial

play08:38

networks as a requirement of the scale.

play08:40

Yeah, that's not necessarily surprising to me.

play08:43

GANs are a specific subclass of deep

play08:46

learning model with specific applications.

play08:49

I'm wondering, maybe a company has

play08:52

used a GAN internally, and then HR knew

play08:54

that and just said, okay, you can be an

play08:55

expert job, but it's pretty specific.

play08:58

Whereas, prompt engineering, even there is a cat!

play09:05

It.

play09:06

I'm keeping this.

play09:07

It's the last bit of train of thought.

play09:08

Oh, prompt engineering is universally useful

play09:10

for pretty much any job, because these models

play09:13

are so phenomenal, and just with a little

play09:15

bit of know how to use them, you You can

play09:18

increase your productivity in so many facets

play09:19

of your work with pretty minimal effort.

play09:22

So it's definitely a more widely applicable and

play09:25

universal skill than something as niche as GANs.

play09:29

What was your kind of intro

play09:32

into prompt engineering?

play09:33

Was it part of your education?

play09:35

Was it existing term around?

play09:39

It was definitely not a

play09:39

formal part of my education.

play09:41

My education, for my master's, focused on

play09:44

theory and machine engineering requirements..

play09:47

Developing models will apply

play09:48

toward that sort of thing.

play09:50

Personally, I first became aware of

play09:52

it, I'd say halfway through my degree.

play09:55

I remember one time I was really struggling

play09:57

with understanding this block of code.

play09:59

And I believe it was a block of code for

play10:02

calculating a prior for a diffusion model that

play10:06

we had to code from scratch in deep learning.

play10:07

So I was really struggling with it, and

play10:09

my wife just said, Can't you just put code

play10:11

in the chatGPT and have it explain it?

play10:13

And I said, I think that works with simple things,

play10:16

but it's, you know, it's pretty high level.

play10:18

It's essentially applied mathematical programming

play10:20

through a pretty complicated mathematical process.

play10:23

I was very skeptical, but I said,

play10:25

what the heck, let me try it.

play10:28

I put it in, I said, hey, explain it.

play10:30

And I was blown away how perfectly and

play10:32

intuitively it was able to take something

play10:34

so technical, something so daunting

play10:36

and unapproachable to someone like me.

play10:38

Who has a solid math background, but

play10:40

it's not super advanced compared to the

play10:43

people really working in this field.

play10:45

And that was the lightbulb moment for

play10:46

me, where I realized, okay, maybe I

play10:48

need to get rid of my preconceived

play10:50

notions about chatGPT can and can't do.

play10:52

And if I believe that it even has a

play10:54

small chance that it can help me with

play10:57

the task, I'm just going to try it.

play10:59

And more often than not, it doesn't always work,

play11:02

but more often than not, it's going to immediately

play11:04

indicate to me that there is some value here.

play11:07

And then you get in the process of, okay,

play11:09

is it not working because it can't do this?

play11:11

Or is it not working because you're not giving

play11:13

it the proper instructions to get it to do it?

play11:16

That was personally my progression.

play11:17

So once I got to that point, I said, okay,

play11:19

now I need to get better at telling chatGPT,

play11:23

which is what I was using, how to help me.

play11:26

And then you start reading stuff.

play11:27

Personally, I've learned a lot from

play11:28

people here at Synthminds because

play11:30

there's incredible prompters here.

play11:32

And they really, just when I thought I was

play11:34

getting good at prompt engineering, look

play11:35

at some of the work that people are doing.

play11:37

And I say, okay, I actually have no idea.

play11:40

These people really know their stuff.

play11:42

They're going way deeper than I thought possible.

play11:44

And then, yeah, I'd say once you get to

play11:46

that point, You've done the easy stuff, but

play11:48

then you really have to learn the techniques

play11:50

for how to extract that extra value.

play11:53

And once you dedicate yourself to learning,

play11:54

even basic prompt engineering can take you way

play11:58

further than blindly typing in the chatGPT.

play12:02

Hi, Video Editing Goddard here.

play12:04

I promise to you, Tanner

play12:05

was not forced to say this.

play12:07

But I think this is a great

play12:09

opportunity to tell you.

play12:10

The second cohort of chatGPT and beyond advanced

play12:13

prompt engineering is starting on July 29th.

play12:17

It is a three week long live course where

play12:19

we teach advanced prompting techniques

play12:22

such as meta prompting and multi role

play12:24

prompting so your chatGPT can prompt itself.

play12:27

The course is project based, so if you are

play12:29

interested, use code GOTTAGO for 10 percent off.

play12:33

I have to say, like, For me, personally,

play12:35

prompting ruined my experience with chatGPT.

play12:38

It completely removed the magic.

play12:41

And I vividly remember that I was

play12:43

asking a question and I, I just

play12:46

thought, okay, so just statistics.

play12:47

So I'm asked this question

play12:49

once and I got this answer.

play12:51

What if I just ask the same question?

play12:53

Do I get the same answer?

play12:54

And I was just curious how

play12:56

same answer can be change.

play12:58

Then I learned about temperature and

play13:00

like top P and stuff like that, but it

play13:01

dawned to me like, okay, so if I'm playing

play13:03

with different settings, but what if I

play13:05

ask the same question a thousand times?

play13:08

And then I saw hallucinations that magically

play13:11

appear, so I plugged, um, plugged API

play13:14

to Google Sheets and I was just like,

play13:16

same question dragged itself down and it

play13:17

was like, oh, so it's just almost luck?

play13:21

What answer I get?

play13:22

Or, can I control what answer I get?

play13:25

And this was like, you know, I started

play13:27

playing with different variables.

play13:29

I remember we were debating very long, is

play13:31

it better to say, act as the copywriter?

play13:35

Or act as blog writer.

play13:37

What gives better results?

play13:39

Anyway, so I'm going on tangent here, but

play13:41

it's always for me fascinating to hear

play13:43

people's experience with prompt engineering.

play13:45

Yeah, but what you were doing sounds

play13:47

like that could be like a master thesis.

play13:48

That's a really interesting way to try to

play13:51

break it down and study it and analyze it.

Rate This

5.0 / 5 (0 votes)

الوسوم ذات الصلة
Ingénierie IAAttentesMachine LearningRéalités professionnellesFormation AcadémiqueIntégration IAModèles PrédictifsGénération IAPrompt EngineeringÉducation TechniqueMarché de l'emploi
هل تحتاج إلى تلخيص باللغة الإنجليزية؟