You think you know what is an NFT? Well, think again...

You are doing it wrong if you think about NFTs as pixelated images of punks, toads, or apes. It is not about the JPEG!

A better mental model for thinking about NFTs πŸ‘‡
Forget the images for now. Owning an NFT means that your wallet address is listed as the owner of a specific digital asset on the blockchain.

Digital assets are organized in collections and an NFT is one specific piece of this collection.

Let's look at an example πŸ‘‡
I own an NFT from the @underfittedio membership collection - a membership card.

You can now check the collection on the blockchain and can see that my wallet address is the owner of token ID 4.

Everybody can check this. Nobody can change it - except me!

How is this useful? πŸ‘‡
@underfittedio I get access to members-only communication channels

In the @underfittedio community, we use Discord for communication (chat, audio, and video). Only people that own a membership NFT get access to the members-only area. Ownership is verified on the blockchain by a bot.

πŸ‘‡
@underfittedio I can vote on decisions for the development of the community

All important decisions in the community are done using voting. To vote, you need to prove that you own a membership NFT. The voting platform (Snapshot) verifies this on the blockchain.

πŸ‘‡
@underfittedio I can participate in the development of community projects

Only members holding the NFT are allowed to participate in building community projects and earn a share of their profit. Again, who is allowed to participate is verified on the blockchain.

πŸ‘‡
@underfittedio Now, let's finally talk about the images!

Every NFT has some additional information (metadata) associated with it. This information is usually not stored directly on the blockchain, but you can get the link to this information from the blockchain (the so-called tokenURI).

πŸ‘‡
@underfittedio The metadata can contain information about different properties on the NFT like a name, description, image, video, 3D model, website... it can really be anything!

Many NFT projects use IPFS as decentralized storage for the metadata.

πŸ‘‡
@underfittedio What do you do with the metadata then? NFT exchanges like OpenSea read the metadata and display it in a nice way on their website, including the image.

So this is how my @underfittedio NFT looks like 🐸

opensea.io/assets/matic/0…

πŸ‘‡
@underfittedio But wait, the image itself doesn't play any role when I'm using the community Discord or voting on Snapshot?

Exactly!

The image, the JPEG is only a *visualization* of the information stored on the blockchain - it doesn't mean anything by itself.

And how could it... πŸ‘‡
@underfittedio Think about this for a second! JPEGs are digital assets, so they can be reproduced infinitely at zero cost. You don't own the JPEG, because everybody can copy it. Or as they say "right-click save-as" the image.

What we really care about is the information stored on the chain πŸ‘‡
@underfittedio The information on the blockchain is what unlocks the utility of an NFT, not the JPEG.

Different NFT projects offer different utilities and you need to decide which is useful for you and which is not when buying an NFT. Additionally, NFTs can be easily traded on exchanges.

πŸ‘‡
@underfittedio NFT utilities are worth a whole thread, but here are some quick examples:

🎫 Membership cards
🎟️ Event tickets (both in the digital and physical world)
🎨 Digital art ownership
πŸ•ΉοΈ Virtual objects in games
🌐 Crypto domain names (like ENS)
πŸŽ–οΈ Achievement badges
...

πŸ‘‡
@underfittedio With this mental model, some of the NFT critiques become very easy to answer.

"I can just right-click and save your NFT"

Yes, you can, but you will only save the image, the metadata. The information on the blockchain is unchanged and this is what gives the whole utility!

πŸ‘‡
@underfittedio "Buying an NFT is not like buying the Mona Lisa, but like buying a receipt for the Mona Lisa"

Sure, digital assets can be copied and distributed at no cost. What I care about is what I can do with the receipt. What if it allowed me to have dinner with Leonardo Da Vinci? πŸ˜€

πŸ‘‡
@underfittedio "I can take your image and mint it as an NFT myself"

Sure, but you can't do this in the original collection, so you still can't use the utility. You can't sell it.

That's like writing a check to send money from my bank account, but signing it yourself - it's worthless.

πŸ‘‡
@underfittedio "It makes no sense to pay thousands of dollars for a JPEG"

It may not matter to you, but it matters to some people to have access to the community and events behind projects like Bored Apes.

It's like buying an expensive watch when a $20 one will do the same job πŸ€·β€β™‚οΈ

πŸ‘‡
@underfittedio "All NFTs are a scam"

There are many scam projects, but there are many genuine ones. The utility of a project may be worthless to you (and this is fine), but this doesn't automatically make it a scam.

πŸ‘‡
@underfittedio "NFTs are used to sell stolen artist's works"

Yes, some people do this. Technology is amoral - it amplifies people's abilities. Unfortunately, there are both good and bad people in the world.

Same thing happens with torrents, fake watches, and Adibas sneakers.

πŸ‘‡
@underfittedio Let's summarize

Think of NFTs as proving you own a piece of a digital collection. This proof gives you access to the utility associated with the particular project.

The image is just a simple visualization of the proof.

It's not about the JPEG!
I'll be sharing more thoughts about crypto, NFTs, and web3 in the future, along with my regular machine learning content.

Follow me @haltakov for an honest, hype-free discussion on these topics πŸ˜ƒ
Yes, this is a good point! NFTs help creators go direct to consumers and get paid.

This is possible for some industries for a long time, but not for all. Music is one example that is currently under a big transformation through NFTs

You can use a centralized database and it will be much more efficient. Using the blockchain is not always the best solution.

However, if you care about avoiding centralization and relying on the mercy of a big tech company, the blockchain is better.

High gas fees are a big problem currently on Ethereum. However, there are many clever people working hard on Ethereum 2.0 where things will drastically improve.

In the meantime, you can use Polygon, Solana, Tezos where gas fees are less than 1 cent πŸ˜‰

This is a good question! Depending on the project, a centralized database may be better because it is way more efficient.

However, then you are at the mercy of a private company that may for example decide to delete your account without any consequences.

Here one example, but you can find much more with other companies as well.

mmos.com/news/ubisoft-d…

β€’ β€’ β€’

Missing some Tweet in this thread? You can try to force a refresh
γ€€

Keep Current with Vladimir Haltakov

Vladimir Haltakov Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @haltakov

21 Dec 21
Things are getting more and more interesting for AI-generated images! 🎨

GLIDE is a new model by @OpenAI that can generate images guided by a text prompt. It is based on a diffusion model instead of the more widely used GAN models.

Some details πŸ‘‡
@OpenAI GLIDE also has the interesting ability to perform inpainting allowing for some interesting usages.

πŸ‘‡
@OpenAI Here is the full paper

arxiv.org/abs/2112.10741

πŸ‘‡
Read 5 tweets
17 Dec 21
How to evaluate your ML model? πŸ“

Your accuracy is 97%, so this is pretty good, right? Right? No! ❌

Just looking at the model accuracy is not enough. Let me tell you about some other metrics:
β–ͺ️ Recall
β–ͺ️ Precision
β–ͺ️ F1 score
β–ͺ️ Confusion matrix

Let's go πŸ‘‡

#RepostFriday
We'll use this example in the whole thread - classifying traffic light colors (e.g. for a self-driving car).

Yellow traffic lights appear much less often, so our dataset may look like this.

This means our model could reach 97% accuracy, by ignoring all 🟑 lights. Not good!

πŸ‘‡
Let's assume now that we trained our model and we get the following predictions.

Do you think this model is good? How can we quantitatively evaluate its performance? How should it be improved?

Let's first discuss the possible error types πŸ‘‡
Read 12 tweets
15 Dec 21
First officially approved Level 3 self-driving system in Germany.

This is significant because it is the first time an autonomous system that takes the *driving responsibility* from the driver is approved for mass production!

europe.autonews.com/automakers/mer…

πŸ‘‡
The main difference between Level 2 and Level 3 systems is that self-driving systems become legally responsible for the actions of the cars when in autonomous mode!

All driver assist systems on the market now (including Tesla) are Level 2 systems.



πŸ‘‡
While Waymo and Cruise have Level 4 systems running as a beta in some cities, there are different challenges putting this tech in consumer vehicles and in cars that don't have a huge sensor rack costing tens of thousands of dollars on the roof.

πŸ‘‡
Read 4 tweets
18 Nov 21
Let's talk about a common problem in ML - imbalanced data βš–οΈ

Imagine we want to detect all pixels belonging to a traffic light from a self-driving car's camera. We train a model with 99.88% performance. Pretty cool, right?

Actually, this model is useless ❌

Let me explain πŸ‘‡
The problem is the data is severely imbalanced - the ratio between traffic light pixels and background pixels is 800:1.

If we don't take any measures, our model will learn to classify each pixel as background giving us 99.88% accuracy. But it's useless!

What can we do? πŸ‘‡
Let me tell you about 3 ways of dealing with imbalanced data:

β–ͺ️ Choose the right evaluation metric
β–ͺ️ Undersampling your dataset
β–ͺ️ Oversampling your dataset
β–ͺ️ Adapting the loss

Let's dive in πŸ‘‡
Read 14 tweets
17 Nov 21
Machine Learning in the Real World 🧠 πŸ€–

ML for real-world applications is much more than designing fancy networks and fine-tuning parameters.

In fact, you will spend most of your time curating a good dataset.

Let's go through the steps of the process together πŸ‘‡
Collect Data πŸ’½

We need to represent the real world as accurately as possible. If some situations are underrepresented we are introducing Sampling Bias.

Sampling Bias is nasty because we'll have high test accuracy, but our model will perform badly when deployed.

πŸ‘‡
Traffic Lights 🚦

Let's build a model to recognize traffic lights for a self-driving car. We need to collect data for different:

β–ͺ️ Lighting conditions
β–ͺ️ Weather conditions
β–ͺ️ Distances and viewpoints
β–ͺ️ Strange variants

And if we sample only 🚦 we won't detect πŸš₯ πŸ€·β€β™‚οΈ

πŸ‘‡
Read 16 tweets
16 Nov 21
Can you detect COVID-19 using Machine Learning? πŸ€”

You have an X-ray or CT scan and the task is to detect if the patient has COVID-19 or not. Sounds doable, right?

None of the 415 ML papers published on the subject in 2020 was usable. Not a single one!

Let's see why πŸ‘‡
Researchers from Cambridge took all papers on the topic published from January to October 2020.

β–ͺ️ 2212 papers
β–ͺ️ 415 after initial screening
β–ͺ️ 62 chosen for detailed analysis
β–ͺ️ 0 with potential for clinical use

healthcare-in-europe.com/en/news/machin…

There are important lessons here πŸ‘‡
Small datasets 🐁

Getting medical data is hard, because of privacy concerns, and at the beginning of the pandemic, there was just not much data in general.

Many papers were using very small datasets often collected from a single hospital - not enough for real evaluation.

πŸ‘‡
Read 10 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

:(