I placed a link on HackerNews and got 300 signups in a week. This post explains what happened.

Product Concept

The product is DbSchemaLibrary. A library of Database Schema.

I'm not sure where this idea originated from. Very fast. Build fast, fail fast, test many ideas, and one will be a hit. I tried it. Let's try it anyway, even though it'll probably fail. I finished The Lean Startup book and wanted to use it.

Database job bores me. Important! I get drowsy working on it. Someone must do it. I remember this happening once. I needed examples at the time. Something similar to Recall (my other project) that I can copy — or at least use as a reference.

Frequently googled. Many tabs open. The results were useless. I raised my hand and agreed to construct the database myself.

It resurfaced. I decided to do something.

Due Diligence

Lean Startup emphasizes validated learning. Everything the startup does should result in learning. I may build something nobody wants otherwise. That's what happened to Recall.

So, I wrote a business plan document. This happens before I code. What am I solving? What is my proposed solution? What is the leap of faith between the problem and solution? Who would be my target audience?

My note:

In my previous project, I did the opposite!

I wrote my expectations after reading the book's advice.

“Failure is a prerequisite to learning. The problem with the notion of shipping a product and then seeing what happens is that you are guaranteed to succeed — at seeing what happens.” — The Lean Startup book

These are successful metrics. If I don't reach them, I'll drop the idea and try another. I didn't understand numbers then. Below are guesses. But it’s a start!

I then wrote the project's What and Why. I'll use this everywhere. Before, I wrote a different pitch each time. I thought certain words would be better. I felt the audience might want something unusual.

Occasionally, this works. I'm unsure if it's a good idea. No stats, just my writing-time opinion. Writing every time is time-consuming and sometimes hazardous. Having a copy saved me duplication.

I can measure and learn from performance.

Last, I identified communities that might demand the product. This became an exercise in creativity.

The MVP

So now it’s time to build.

A MVP can test my assumptions. Business may learn from it. Not low-quality. We should learn from the tiniest thing.

I like the example of how Dropbox did theirs. They assumed that if the product works, people will utilize it. How can this be tested without a quality product? They made a movie demonstrating the software's functionality. Who knows how much functionality existed?

So I tested my biggest assumption. Users want schema references. How can I test if users want to reference another schema? I'd love this. Recall taught me that wanting something doesn't mean others do.

I made an email-collection landing page. Describe it briefly. Reference library. Each email sender wants a reference. They're interested in the product. Few other reasons exist.

Header and footer were skipped. No name or logo. DbSchemaLibrary is a name I thought of after the fact. 5-minute logo. I expected a flop. Recall has no users after months of labor. What could happen to a 2-day project?

I didn't compromise learning validation. How many visitors sign up? To draw a conclusion, I must track these results.

Posting Time

Now that the job is done, gauge interest. The next morning, I posted on all my channels. I didn't want to be spammy, therefore it required more time.

I made sure each channel had at least one fan of this product. I also answer people's inquiries in the channel.

My list stinks. Several channels wouldn't work. The product's target market isn't there. Posting there would waste our time. This taught me to create marketing channels depending on my persona.

Statistics! What actually happened

My favorite part! 23 channels received the link.

I stopped posting to Discord despite its high conversion rate. I eliminated some channels because they didn't fit. According to the numbers, some users like it. Most users think it's spam.

I was skeptical. And 12 people viewed it.

I didn't expect much attention on a startup subreddit. I'll likely examine Reddit further in the future. As I have enough info, I didn't post much. Time for the next validated learning

No comment. The post had few views, therefore the numbers are low.

The targeted people come next.

I'm a Toptal freelancer. There's a member-only Slack channel. Most people can't use this marketing channel, but you should! It's not as spectacular as discord's 27% conversion rate. But I think the users here are better.

I don’t really have a following anywhere so this isn’t something I can leverage.

The best yet. 10% is converted. With more data, I expect to attain a 10% conversion rate from other channels. Stable number.

This number required some work. Did you know that people use many different clients to read HN?

Unknowns

Untrackable views and signups abound. 1136 views and 135 signups are untraceable. It's 11%. I bet much of that came from Hackernews.

Overall Statistics

The 7-day signup-to-visit ratio was 17%. (Hourly data points)

First-day percentages were lower, which is noteworthy. Initially, it was little above 10%. The HN post started getting views then.

When traffic drops, the number reaches just around 20%. More individuals are interested in the connection. hn.algolia.com sent 2 visitors. This means people are searching and finding my post.

Interesting discoveries

1. HN post struggled till the US woke up.

11am UTC. After an hour, it lost popularity. It seemed over. 7 signups converted 13%. Not amazing, but I would've thought ahead.

After 4pm UTC, traffic grew again. 4pm UTC is 9am PDT. US awakened. 10am PDT saw 512 views.

2. The product was highlighted in a newsletter.

I found Revue references when gathering data. Newsletter platform. Someone posted the newsletter link. 37 views and 3 registrations.

3. HN numbers are extremely reliable

I don't have a time-lapse graph (yet). The statistics were constant all day.

• 2717 views later 272 new users, or 10.1%

• With 293 signups at 2856 views, 10.25%

• At 306 signups at 2965 views, 10.32%

Learnings

1. My initial estimations were wildly inaccurate

I wrote 30% conversion. Reading some articles, looks like 10% is a good number to aim for.

2. Paying attention to what matters rather than vain metrics

The Lean Startup discourages vanity metrics. Feel-good metrics that don't measure growth or traction. Considering the proportion instead of the total visitors made me realize there was something here.

What’s next?

There are lots of work to do. Data aggregation, display, website development, marketing, legal issues. Fun! It's satisfying to solve an issue rather than investigate its cause.

In the meantime, I’ve already written the first project update in another post. Continue reading it if you’d like to know more about the project itself! Shifting from Quantity to Quality — DbSchemaLibrary

New social media producers' biggest error

The Treadmill of Social Media Content

New creators focus on the wrong platforms.

They post to Instagram, Twitter, TikTok, etc.

They create daily material, but it's never enough for social media algorithms.

Creators recognize they're on a content creation treadmill.

They have to keep publishing content daily just to stay on the algorithm’s good side and avoid losing the audience they’ve built on the platform.

This is exhausting and unsustainable, causing creator burnout.

They focus on short-lived platforms, which is an issue.

Comparing low- and high-return social media platforms

Social media networks are great for reaching new audiences.

Their algorithm is meant to viralize material.

Social media can use you for their aims if you're not careful.

To master social media, focus on the right platforms.

To do this, we must differentiate low-ROI and high-ROI platforms:

Low ROI platforms are ones where content has a short lifespan. High ROI platforms are ones where content has a longer lifespan.

A tweet may be shown for 12 days. If you write an article or blog post, it could get visitors for 23 years.

ROI is drastically different.

New creators have limited time and high learning curves.

Nothing is possible.

First create content for high-return platforms.

ROI for social media platforms

Here are high-return platforms:

1. Your Blog - A single blog article can rank and attract a ton of targeted traffic for a very long time thanks to the power of SEO.

2. YouTube - YouTube has a reputation for showing search results or sidebar recommendations for videos uploaded 23 years ago. A superb video you make may receive views for a number of years.

3. Medium - A platform dedicated to excellent writing is called Medium. When you write an article about a subject that never goes out of style, you're building a digital asset that can drive visitors indefinitely.

These high ROI platforms let you generate content once and get visitors for years.

This contrasts with low ROI platforms:

1. Twitter

2. Instagram

3. TikTok

4. LinkedIn

5. Facebook

The posts you publish on these networks have a 23-day lifetime. Instagram Reels and TikToks are exceptions since viral content can last months.

If you want to make content creation sustainable and enjoyable, you must focus the majority of your efforts on creating high ROI content first. You can then use the magic of repurposing content to publish content to the lower ROI platforms to increase your reach and exposure.

How To Use Your Content Again

So, you’ve decided to focus on the high ROI platforms.

Great!

You've published an article or a YouTube video.

You worked hard on it.

Now you have fresh stuff.

What now?

If you are not repurposing each piece of content for multiple platforms, you are throwing away your time and efforts.

You've created fantastic material, so why not distribute it across platforms?

Repurposing Content Step-by-Step

For me, it's writing a blog article, but you might start with a video or podcast.

The premise is the same regardless of the medium.

Start by creating content for a high ROI platform (YouTube, Blog Post, Medium). Then, repurpose, edit, and repost it to the lower ROI platforms.

Here's how to repurpose pillar material for other platforms:

1. Post the article on your blog.

2. Put your piece on Medium (use the canonical link to point to your blog as the source for SEO)

3. Create a video and upload it to YouTube using the talking points from the article.

4. Rewrite the piece a little, then post it to LinkedIn.

5. Change the article's format to a Thread and share it on Twitter.

6. Find a few quick quotes throughout the article, then use them in tweets or Instagram quote posts.

7. Create a carousel for Instagram and LinkedIn using screenshots from the Twitter Thread.

8. Go through your film and select a few valuable 30-second segments. Share them on LinkedIn, Facebook, Twitter, TikTok, YouTube Shorts, and Instagram Reels.

9. Your video's audio can be taken out and uploaded as a podcast episode.

If you (or your team) achieve all this, you'll have 20-30 pieces of social media content.

If you're just starting, I wouldn't advocate doing all of this at once.

Instead, focus on a few platforms with this method.

You can outsource this as your company expands. (If you'd want to learn more about content repurposing, contact me.)

You may focus on relevant work while someone else grows your social media on autopilot.

You develop high-ROI pillar content, and it's automatically chopped up and posted on social media.

This lets you use social media algorithms without getting sucked in.

Thanks for reading!

No Paid Ads Required

I had an online store. After a year of running the company alongside my 9-to-5, I made enough to resign.

That day was amazing.

This Instagram marketing plan helped the store succeed.

How did I increase my sales to five figures a month without using any paid advertising?

I used customer event marketing.

I'm not sure this term exists. I invented it to describe what I was doing.

Instagram word-of-mouth, fan engagement, and interaction drove sales.

If a customer liked or disliked a product, the buzz would drive attention to the store.

I used customer-based events to increase engagement and store sales.

Success!

Here are the weekly Instagram customer events I coordinated while running my business:

• Be the Buyer Days

• Flash sales

• Mystery boxes

Be the Buyer Days: How do they work?

Be the Buyer Days are exactly that.

You choose a day to share stock selections with social media followers.

This is an easy approach to engaging customers and getting fans enthusiastic about new releases.

First, pick a handful of items you’re considering ordering. I’d usually pick around 3 for Be the Buyer Day.

Then I'd poll the crowd on Instagram to vote on their favorites.

This was before Instagram stories, polls, and all the other cool features Instagram offers today. I think using these tools now would make this event even better.

I'd ask customers their favorite back then.

The growing comments excited customers.

Then I'd declare the winner, acquire the products, and start selling it.

How do flash sales work?

I mostly ran flash sales.

You choose a limited number of itemsdd for a few-hour sale.

We wanted most sales to result in sold-out items.

When an item sells out, it contributes to the sensation of scarcity and can inspire customers to visit your store to buy a comparable product, join your email list, become a fan, etc.

We hoped they'd act quickly.

I'd hold flash deals twice a week, which generated scarcity and boosted sales.

The store had a few thousand Instagram followers when I started flash deals.

Each flash sale item would make $400 to $600.

$400 x 3= $1,200

That's $1,200 on social media!

Twice a week, you'll make roughly $10K a month from Instagram.

$1,200/day x 8 events/month=$9,600

Flash sales did great.

We held weekly flash deals and sent social media and email reminders. That’s about it!

How are mystery boxes put together?

All you do is package a box of store products and sell it as a mystery box on TikTok or retail websites.

A $100 mystery box would cost $30.

You're discounting high-value boxes.

This is a clever approach to get rid of excess inventory and makes customers happy.

It worked!

Be the Buyer Days, flash deals, and mystery boxes helped build my company without paid advertisements.

All companies can use customer event marketing. Involving customers and providing an engaging environment can boost sales.

Try it!

After a decade in large tech, I know how software engineers were interviewed. I've seen outstanding engineers fail interviews because their responses were too vague.

So I wrote Nail A Coding Interview: Six-Step Mental Framework. Give candidates a mental framework for coding questions; help organizations better prepare candidates so they can calibrate traits.

Recently, I sold more than 100 books, something I never expected.

In this essay, I'll describe my publication journey, which included self-doubt and little triumphs. I hope this helps if you want to publish.

It was originally a Medium post.

How did I know to develop a coding interview book? Years ago, I posted on Medium.

Six steps to ace a coding interview Inhale. blog.devgenius.io

This story got a lot of attention and still gets a lot of daily traffic. It indicates this domain's value.

Converted the Medium article into an ebook

The Medium post contains strong bullet points, but it is missing the “flesh”. How to use these strategies in coding interviews, for example. I filled in the blanks and made a book.

I made the book cover for free. It's tidy.

Shared the article with my close friends on my social network WeChat.

I shared the book on Wechat's Friend Circle (朋友圈) after publishing it on Gumroad. Many friends enjoyed my post. It definitely triggered endorphins.

In Friend Circle, I presented a 100% off voucher. No one downloaded the book. Endorphins made my heart sink.

Several days later, my Apple Watch received a Gumroad notification. A friend downloaded it. I majored in finance, he subsequently said. My brother-in-law can get it? He downloaded it to cheer me up.

I liked him, but was disappointed that he didn't read it.

The Tipping Point: Reddit's Free Giving

I trusted the book. It's based on years of interviewing. I felt it might help job-hunting college students. If nobody wants it, it can still have value.

I posted the book's link on /r/leetcode. I told them to DM me for a free promo code.

Momentum shifted everything. Gumroad notifications kept coming when I was out with family. Following orders.

As promised, I sent DMs a promo code. Some consumers ordered without asking for a promo code. Some readers finished the book and posted reviews.

My book was finally on track.

A 5-Star Review, plus More

A reader afterwards DMed me and inquired if I had another book on system design interviewing. I said that was a good idea, but I didn't have one. If you write one, I'll be your first reader.

Later, I asked for a book review. Yes, but how? That's when I learned readers' reviews weren't easy. I built up an email pipeline to solicit customer reviews. Since then, I've gained credibility through ratings.

Learnings

I wouldn't have gotten 100 if I gave up when none of my pals downloaded. Here are some lessons.

• Your friends are your allies, but they are not your clients.

• Be present where your clients are

• Request ratings and testimonials

• gain credibility gradually

I did it, so can you. Follow me on Twitter @imgracehuang for my publishing and entrepreneurship adventure.

Reddit's NFT market is alive and well.

NFT owners outnumber OpenSea on Reddit.

Reddit NFTs flip in OpenSea in days:

Fast-selling.

NFT sales will make Reddit's current communities more engaged.

I don't think NFTs will affect existing groups, but they will build hype for people to acquire them.

The first season of Collectibles is unique, but many missed the first season.

Second-season NFTs are less likely to be sold for a higher price than first-season ones.

If you use Reddit, it's fun to own NFTs.

Deep learning-based generative art is being researched. As usual, self-learning is better. Some models, like OpenAI's DALL-E 2, require registration and can only be used online, but others can be used locally, which is usually more enjoyable for curious users. I'll demonstrate the Stable Diffusion model's operation on a standard PC.

Let’s get started.

What It Does

Stable Diffusion uses numerous components:

• A generative model trained to produce images is called a diffusion model. The model is incrementally improving the starting data, which is only random noise. The model has an image, and while it is being trained, the reversed process is being used to add noise to the image. Being able to reverse this procedure and create images from noise is where the true magic is (more details and samples can be found in the paper).

• An internal compressed representation of a latent diffusion model, which may be altered to produce the desired images, is used (more details can be found in the paper). The capacity to fine-tune the generation process is essential because producing pictures at random is not very attractive (as we can see, for instance, in Generative Adversarial Networks).

• A neural network model called CLIP (Contrastive Language-Image Pre-training) is used to translate natural language prompts into vector representations. This model, which was trained on 400,000,000 image-text pairs, enables the transformation of a text prompt into a latent space for the diffusion model in the scenario of stable diffusion (more details in that paper).

This figure shows all data flow:

The weights file size for Stable Diffusion model v1 is 4 GB and v2 is 5 GB, making the model quite huge. The v1 model was trained on 256x256 and 512x512 LAION-5B pictures on a 4,000 GPU cluster using over 150.000 NVIDIA A100 GPU hours. The open-source pre-trained model is helpful for us. And we will.

Install

Before utilizing the Python sources for Stable Diffusion v1 on GitHub, we must install Miniconda (assuming Git and Python are already installed):

wget https://repo.anaconda.com/miniconda/Miniconda3-py39_4.12.0-Linux-x86_64.sh
chmod +x Miniconda3-py39_4.12.0-Linux-x86_64.sh
./Miniconda3-py39_4.12.0-Linux-x86_64.sh
conda update -n base -c defaults conda

Install the source and prepare the environment:

git clone https://github.com/CompVis/stable-diffusion
cd stable-diffusion
conda env create -f environment.yaml
conda activate ldm
pip3 install transformers --upgrade

Download the pre-trained model weights next. HiggingFace has the newest checkpoint sd-v14.ckpt (a download is free but registration is required). Put the file in the project folder and have fun:

python3 scripts/txt2img.py --prompt "hello world" --plms --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

Almost. The installation is complete for happy users of current GPUs with 12 GB or more VRAM. RuntimeError: CUDA out of memory will occur otherwise. Two solutions exist.

Running the optimized version

Try optimizing first. After cloning the repository and enabling the environment (as previously), we can run the command:

python3 optimizedSD/optimized_txt2img.py --prompt "hello world" --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

Stable Diffusion worked on my visual card with 8 GB RAM (alas, I did not behave well enough to get NVIDIA A100 for Christmas, so 8 GB GPU is the maximum I have;).

Running Stable Diffusion without GPU

If the GPU does not have enough RAM or is not CUDA-compatible, running the code on a CPU will be 20x slower but better than nothing. This unauthorized CPU-only branch from GitHub is easiest to obtain. We may easily edit the source code to use the latest version. It's strange that a pull request for that was made six months ago and still hasn't been approved, as the changes are simple. Readers can finish in 5 minutes:

• Replace if attr.device!= torch.device(cuda) with if attr.device!= torch.device(cuda) and torch.cuda.is available at line 20 of ldm/models/diffusion/ddim.py ().

• Replace if attr.device!= torch.device(cuda) with if attr.device!= torch.device(cuda) and torch.cuda.is available in line 20 of ldm/models/diffusion/plms.py ().

• Replace device=cuda in lines 38, 55, 83, and 142 of ldm/modules/encoders/modules.py with device=cuda if torch.cuda.is available(), otherwise cpu.

• Replace model.cuda() in scripts/txt2img.py line 28 and scripts/img2img.py line 43 with if torch.cuda.is available(): model.cuda ().

Run the script again.

Testing

Test the model. Text-to-image is the first choice. Test the command line example again:

python3 scripts/txt2img.py --prompt "hello world" --plms --ckpt sd-v1-4.ckpt --skip_grid --n_samples 1

The slow generation takes 10 seconds on a GPU and 10 minutes on a CPU. Final image:

Hello world is dull and abstract. Try a brush-wielding hamster. Why? Because we can, and it's not as insane as Napoleon's cat. Another image:

Generating an image from a text prompt and another image is interesting. I made this picture in two minutes using the image editor (sorry, drawing wasn't my strong suit):

I can create an image from this drawing:

python3 scripts/img2img.py --prompt "A bird is sitting on a tree branch" --ckpt sd-v1-4.ckpt --init-img bird.png --strength 0.8

It was far better than my initial drawing:

I hope readers understand and experiment.

Stable Diffusion UI

Developers love the command line, but regular users may struggle. Stable Diffusion UI projects simplify image generation and installation. Simple usage:

• Unpack the ZIP after downloading it from https://github.com/cmdr2/stable-diffusion-ui/releases. Linux and Windows are compatible with Stable Diffusion UI (sorry for Mac users, but those machines are not well-suitable for heavy machine learning tasks anyway;).

• Start the script.

Done. The web browser UI makes configuring various Stable Diffusion features (upscaling, filtering, etc.) easy:

V2.1 of Stable Diffusion

I noticed the notification about releasing version 2.1 while writing this essay, and it was intriguing to test it. First, compare version 2 to version 1:

• alternative text encoding. The Contrastive LanguageImage Pre-training (CLIP) deep learning model, which was trained on a significant number of text-image pairs, is used in Stable Diffusion 1. The open-source CLIP implementation used in Stable Diffusion 2 is called OpenCLIP. It is difficult to determine whether there have been any technical advancements or if legal concerns were the main focus. However, because the training datasets for the two text encoders were different, the output results from V1 and V2 will differ for the identical text prompts.

• a new depth model that may be used to the output of image-to-image generation.

• a revolutionary upscaling technique that can quadruple the resolution of an image.

• Generally higher resolution Stable Diffusion 2 has the ability to produce both 512x512 and 768x768 pictures.

The Hugging Face website offers a free online demo of Stable Diffusion 2.1 for code testing. The process is the same as for version 1.4. Download a fresh version and activate the environment:

conda deactivate  
conda env remove -n ldm  # Use this if version 1 was previously installed
git clone https://github.com/Stability-AI/stablediffusion
cd stablediffusion
conda env create -f environment.yaml
conda activate ldm

Hugging Face offers a new weights ckpt file.

The Out of memory error prevented me from running this version on my 8 GB GPU. Version 2.1 fails on CPUs with the slow conv2d cpu not implemented for Half error (according to this GitHub issue, the CPU support for this algorithm and data type will not be added). The model can be modified from half to full precision (float16 instead of float32), however it doesn't make sense since v1 runs up to 10 minutes on the CPU and v2.1 should be much slower. The online demo results are visible. The same hamster painting with a brush prompt yielded this result:

It looks different from v1, but it functions and has a higher resolution.

The superresolution.py script can run the 4x Stable Diffusion upscaler locally (the x4-upscaler-ema.ckpt weights file should be in the same folder):

python3 scripts/gradio/superresolution.py configs/stable-diffusion/x4-upscaling.yaml x4-upscaler-ema.ckpt

This code allows the web browser UI to select the image to upscale:

The copy-paste strategy may explain why the upscaler needs a text prompt (and the Hugging Face code snippet does not have any text input as well). I got a GPU out of memory error again, although CUDA can be disabled like v1. However, processing an image for more than two hours is unlikely:

Stable Diffusion Limitations

When we use the model, it's fun to see what it can and can't do. Generative models produce abstract visuals but not photorealistic ones. This fundamentally limits The generative neural network was trained on text and image pairs, but humans have a lot of background knowledge about the world. The neural network model knows nothing. If someone asks me to draw a Chinese text, I can draw something that looks like Chinese but is actually gibberish because I never learnt it. Generative AI does too! Humans can learn new languages, but the Stable Diffusion AI model includes only language and image decoder brain components. For instance, the Stable Diffusion model will pull NO WAR banner-bearers like this:

V1:

V2.1:

The shot shows text, although the model never learned to read or write. The model's string tokenizer automatically converts letters to lowercase before generating the image, so typing NO WAR banner or no war banner is the same.

I can also ask the model to draw a gorgeous woman:

V1:

V2.1:

The first image is gorgeous but physically incorrect. A second one is better, although it has an Uncanny valley feel. BTW, v2 has a lifehack to add a negative prompt and define what we don't want on the image. Readers might try adding horrible anatomy to the gorgeous woman request.

If we ask for a cartoon attractive woman, the results are nice, but accuracy doesn't matter:

V1:

V2.1:

Another example: I ordered a model to sketch a mouse, which looks beautiful but has too many legs, ears, and fingers:

V1:

V2.1: improved but not perfect.

V1 produces a fun cartoon flying mouse if I want something more abstract:

I tried multiple times with V2.1 but only received this:

The image is OK, but the first version is closer to the request.

Stable Diffusion struggles to draw letters, fingers, etc. However, abstract images yield interesting outcomes. A rural landscape with a modern metropolis in the background turned out well:

V1:

V2.1:

Generative models help make paintings too (at least, abstract ones). I searched Google Image Search for modern art painting to see works by real artists, and this was the first image:

I typed "abstract oil painting of people dancing" and got this:

V1:

V2.1:

It's a different style, but I don't think the AI-generated graphics are worse than the human-drawn ones.

The AI model cannot think like humans. It thinks nothing. A stable diffusion model is a billion-parameter matrix trained on millions of text-image pairs. I input "robot is creating a picture with a pen" to create an image for this post. Humans understand requests immediately. I tried Stable Diffusion multiple times and got this:

This great artwork has a pen, robot, and sketch, however it was not asked. Maybe it was because the tokenizer deleted is and a words from a statement, but I tried other requests such robot painting picture with pen without success. It's harder to prompt a model than a person.

I hope Stable Diffusion's general effects are evident. Despite its limitations, it can produce beautiful photographs in some settings. Readers who want to use Stable Diffusion results should be warned. Source code examination demonstrates that Stable Diffusion images feature a concealed watermark (text StableDiffusionV1 and SDV2) encoded using the invisible-watermark Python package. It's not a secret, because the official Stable Diffusion repository's test watermark.py file contains a decoding snippet. The put watermark line in the txt2img.py source code can be removed if desired. I didn't discover this watermark on photographs made by the online Hugging Face demo. Maybe I did something incorrectly (but maybe they are just not using the txt2img script on their backend at all).

Conclusion

The Stable Diffusion model was fascinating. As I mentioned before, trying something yourself is always better than taking someone else's word, so I encourage readers to do the same (including this article as well;).

Is Generative AI a game-changer? My humble experience tells me:

• I think that place has a lot of potential. For designers and artists, generative AI can be a truly useful and innovative tool. Unfortunately, it can also pose a threat to some of them since if users can enter a text field to obtain a picture or a website logo in a matter of clicks, why would they pay more to a different party? Is it possible right now? unquestionably not yet. Images still have a very poor quality and are erroneous in minute details. And after viewing the image of the stunning woman above, models and fashion photographers may also unwind because it is highly unlikely that AI will replace them in the upcoming years.

• Today, generative AI is still in its infancy. Even 768x768 images are considered to be of a high resolution when using neural networks, which are computationally highly expensive. There isn't an AI model that can generate high-resolution photographs natively without upscaling or other methods, at least not as of the time this article was written, but it will happen eventually.

• It is still a challenge to accurately represent knowledge in neural networks (information like how many legs a cat has or the year Napoleon was born). Consequently, AI models struggle to create photorealistic photos, at least where little details are important (on the other side, when I searched Google for modern art paintings, the results are often even worse;).

• When compared to the carefully chosen images from official web pages or YouTube reviews, the average output quality of a Stable Diffusion generation process is actually less attractive because to its high degree of randomness. When using the same technique on their own, consumers will theoretically only view those images as 1% of the results.

Anyway, it's exciting to witness this area's advancement, especially because the project is open source. Google's Imagen and DALL-E 2 can also produce remarkable findings. It will be interesting to see how they progress.