Christopher Janz of Point Nine Capital created the SaaS napkin in 2016. This post shows how founders have raised cash in the last 6 years. View raw data.

Round size

Unsurprisingly, round sizes have expanded and will taper down in 2022. In 2016, pre-seed rounds were $200k to $500k; currently, they're $1-$2m. Despite the macroeconomic scenario, Series A have expanded from $3m to $12m in 2016 to $6m and $18m in 2022.

Valuation

There are hints that valuations are rebounding this year. Pre-seed valuations in 2022 are $12m from $3m in 2016, and Series B prices are $270m from $100m in 2016.

Compared to public SaaS multiples, Series B valuations more closely reflect the market, but Seed and Series A prices seem to be inflated regardless of the market.

I'd like to know how each annual cohort performed for investors, based on the year they invested and the valuations. I can't access this information.

ARR

Seed firms' ARR forecasts have risen from $0 to $0.6m to $0 to $1m. 2016 expected $1.2m to $3m, 2021 $0.5m to $4m, and this year $0.5m to $2.5m, suggesting that Series A firms may raise with less ARR today. Series B minutes fell from $4.2m to $3m.

Capitalization Rate

2022 is the year that VCs start discussing capital efficiency in portfolio meetings. Given the economic shift in the markets and the stealthy VC meltdown, it's not surprising. Christopher Janz added capital efficiency to the SaaS Napkin as a new statistic for Series A (3.5x) and Series B. (2.5x). Your investors must live under a rock if they haven't asked about capital efficiency. If you're unsure:

The Capital Efficiency Ratio is the ratio of how much a company has spent growing revenue and how much they’re receiving in return. It is the broadest measure of company effectiveness in generating ARR

What next?

No one knows what's next, including me. All startup and growing enterprises around me are tightening their belts and extending their runways in anticipation of a difficult fundraising ride. If you're wanting to raise money but can wait, wait till the market is more stable and access to money is easier.

You’ll be surprised; I sure was

Blogging and writing online as a side income has now been around for a significant amount of time. Nowadays, it is a continuously rising moneymaker for prospective writers, with several writing platforms existing online. At the top of the list are Medium, Vocal Media, Newsbreak, and the biggest one of them, Quora, with 300 million active users.

Quora, unlike Medium, is a question-and-answer format platform. On Medium you are permitted to write what you want, while on Quora, you answer questions on topics that you have expertise about. Quora, like Medium, now compensates its authors for the answers they provide in comparison to the previous, in which you had to be admitted to the partner program and were paid to ask questions.

Quora just recently went live with this new partner program, Quora Plus, and the way it works is that it is a subscription for $5 a month which provides you access to metered/monetized stories, in turn compensating the writers for part of that subscription for their answers.

I too on Quora have found a lot of success on the platform, gaining 23 Million Content Views, and 300,000 followers for my space, which is kind of the Quora equivalent of a Medium article. The way in which I was able to do this was entirely thanks to a hack that I uncovered to the Quora algorithm.

In this article, I plan on discussing how much money I received from 23 million content views on Quora, and I bet you’ll be shocked; I know I was.

A Brief Explanation of How I Got 23 Million Views and How You Can Do It Too

On Quora, everything in terms of obtaining views is about finding the proper question, which I only understood quite late into the game. I published my first response in 2019 but never actually wrote on Quora until the summer of 2020, and about a month into posting consistently I found out how to find the perfect question. Here’s how:

The Process

Go to your Home Page and start scrolling… While browsing, check for the following things…

1. Answers from people you follow or your followers.

2. Advertisements

These two things are the two things you want to ignore, you don’t want to answer those questions or look at the ads. You should now be left with a couple of recommended answers. To discover which recommended answer is the best to answer as well, look at these three important aspects.

1. Date of the answer: Was it in the past few days, preferably 2–3 days, even better, past 24 hours?

2. Views: Are they in the ten thousands or hundred thousands?

3. Upvotes: Are they in the hundreds or thousands?

Now, choose an answer to a question which you think you could answer as well that satisfies the requirements above. Once you click on it, as all answers on Quora works, it will redirect you to the page for that question, in which you will have to select once again if you should answer the question.

1. Amount of answers: How many responses are there to the given question? This tells you how much competition you have. My rule is beyond 25 answers, you shouldn’t answer, but you can change it anyway you’d like.

2. Answerers: Who did the answering for the question? If the question includes a bunch of renowned, extremely well-known people on Quora, there’s a good possibility your essay is going to get drowned out.

3. Views: Check for a constant quantity of high views on each answer for the question; this is what will guarantee that your answer gets a lot of views!

The Income Reveal! How Much I Made From 23 Million Content Views

DRUM ROLL, PLEASE!

8.97 USD. Yes, not even ten dollars, not even nine. Just eight dollars and ninety-seven cents.

Possible Reasons for My Low Earnings

• Quora Plus and the answering partner program are newer than my Quora views.

• Few people use Quora+, therefore revenues are low.

• I haven't been writing much on Quora, so I'm only making money from old answers and a handful since Quora Plus launched.

• Quora + pays poorly...

Should You Try Quora and Quora For Money?

My answer depends on your needs. I never got invited to Quora's question partner program due to my late start, but other writers have made hundreds. Due to Quora's new and competitive answering partner program, you may not make much money.

If you want a fun writing community, try Quora. Quora was fun when I only made money from my space. Quora +'s paywalls and new contributors eager to make money have made the platform less fun for me.

This article is a summary to save you time. You can read my full, more detailed article, here.

Y Combinators Winter 2022 Demo Day continues the trend of more startups engaging in accelerator Demo Days. Our team evaluated almost 400 projects in Y Combinator's ninth year.

After Winter 2021 Demo Day, we noticed a hurry pushing shorter rounds, inflated valuations, and larger batches.

Despite the batch size, this event's behavior showed a return to normalcy. Our observations show that investors evaluate and fund businesses more carefully. Unlike previous years, more YC businesses gave investors with data rooms and thorough pitch decks in addition to valuation data before Demo Day.

Demo Day pitches were virtual and fast-paced, limiting unplanned meetings. Investors had more time and information to do their due research before meeting founders. Our staff has more time to study diverse areas and engage with interesting entrepreneurs and founders.

This was one of the most regionally diversified YC cohorts to date. This year's Winter Demo Day startups showed some interesting tendencies.

Trends and Industries to Watch Before Demo Day

Demo day events at any accelerator show how investment competition is influencing startups. As startups swiftly become scale-ups and big success stories in fintech, e-commerce, healthcare, and other competitive industries, entrepreneurs and early-stage investors feel pressure to scale quickly and turn a notion into actual innovation.

Too much eagerness can lead founders to focus on market growth and team experience instead of solid concepts, technical expertise, and market validation. Last year, YC Winter Demo Day funding cycles ended too quickly and valuations were unrealistically high.

Scrum Ventures observed a longer funding cycle this year compared to last year's Demo Day. While that seems promising, many factors could be contributing to change, including:

• Market patterns are changing and the economy is becoming worse.

• the industries that investors are thinking about.

• Individual differences between each event batch and the particular businesses and entrepreneurs taking part

The Winter 2022 Batch's Trends

Each year, we also wish to examine trends among early-stage firms and YC event participants. More international startups than ever were anticipated to present at Demo Day.

Less than 50% of demo day startups were from the U.S. For the S21 batch, firms from outside the US were most likely in Latin America or Europe, however this year's batch saw a large surge in startups situated in Asia and Africa.

YC Startup Directory

163 out of 399 startups were B2B software and services companies. Financial, healthcare, and consumer startups were common.

Our team doesn't plan to attend every pitch or speak with every startup's founders or team members. Let's look at cleantech, Web3, and health and wellness startup trends.

Our Opinions Following Conversations with 87 Startups at Demo Day

In the lead-up to Demo Day, we spoke with 87 of the 125 startups going. Compared to B2C enterprises, B2B startups had higher average valuations. A few outliers with high valuations pushed B2B and B2C means above the YC-wide mean and median.

Many of these startups develop business and technology solutions we've previously covered. We've seen API, EdTech, creative platforms, and cybersecurity remain strong and increase each year.

While these persistent tendencies influenced the startups Scrum Ventures looked at and the founders we interacted with on Demo Day, new trends required more research and preparation. Let's examine cleantech, Web3, and health and wellness startups.

Hardware and software that is green

Cleantech enterprises demand varying amounts of funding for hardware and software. Although the same overarching trend is fueling the growth of firms in this category, each subgroup has its own strategy and technique for investigation and identifying successful investments.

Many cleantech startups we spoke to during the YC event are focused on helping industrial operations decrease or recycle carbon emissions.

• Carbon Crusher: Creating carbon negative roads

• Phase Biolabs: Turning carbon emissions into carbon negative products and carbon neutral e-fuels

• Seabound: Capturing carbon dioxide emissions from ships

• Fleetzero: Creating electric cargo ships

• Impossible Mining: Sustainable seabed mining

• Beyond Aero: Creating zero-emission private aircraft

• Verdn: Helping businesses automatically embed environmental pledges for product and service offerings, boost customer engagement

• AeonCharge: Allowing electric vehicle (EV) drivers to more easily locate and pay for EV charging stations

• Phoenix Hydrogen: Offering a hydrogen marketplace and a connected hydrogen hub platform to connect supply and demand for hydrogen fuel and simplify hub planning and partner program expansion

• Aklimate: Allowing businesses to measure and reduce their supply chain’s environmental impact

• Pina Earth: Certifying and tracking the progress of businesses’ forestry projects

• AirMyne: Developing machines that can reverse emissions by removing carbon dioxide from the air

• Unravel Carbon: Software for enterprises to track and reduce their carbon emissions

Web3: NFTs, the metaverse, and cryptocurrency

Web3 technologies handle a wide range of business issues. This category includes companies employing blockchain technology to disrupt entertainment, finance, cybersecurity, and software development.

Many of these startups overlap with YC's FinTech trend. Despite this, B2C and B2B enterprises were evenly represented in Web3. We examined:

• Stablegains: Offering consistent interest on cash balance from the decentralized finance (DeFi) market

• LiquiFi: Simplifying token management with automated vesting contracts, tax reporting, and scheduling. For companies, investors, and finance & accounting

• NFTScoring: An NFT trading platform

• CypherD Wallet: A multichain wallet for crypto and NFTs with a non-custodial crypto debit card that instantly converts coins to USD

• Remi Labs: Allowing businesses to more easily create NFT collections that serve as access to products, memberships, events, and more

• Cashmere: A crypto wallet for Web3 startups to collaboratively manage funds

• Chaingrep: An API that makes blockchain data human-readable and tokens searchable

• Courtyard: A platform for securely storing physical assets and creating 3D representations as NFTs

• Arda: “Banking as a Service for DeFi,” an API that FinTech companies can use to embed DeFi products into their platforms

• earnJARVIS: A premium cryptocurrency management platform, allowing users to create long-term portfolios

• Mysterious: Creating community-specific experiences for Web3 Discords

• Winter: An embeddable widget that allows businesses to sell NFTs to users purchasing with a credit card or bank transaction

• SimpleHash: An API for NFT data that provides compatibility across blockchains, standardized metadata, accurate transaction info, and simple integration

• Lifecast: Tools that address motion sickness issues for 3D VR video

• Gym Class: Virtual reality (VR) multiplayer basketball video game

• WorldQL: An asset API that allows NFT creators to specify multiple in-game interpretations of their assets, increasing their value

• Bonsai Desk: A software development kit (SDK) for 3D analytics

• Campfire: Supporting virtual social experiences for remote teams

• Unai: A virtual headset and Visual World experience

• Vimmerse: Allowing creators to more easily create immersive 3D experiences

Fitness and health

Scrum Ventures encountered fewer health and wellness startup founders than Web3 and Cleantech. The types of challenges these organizations solve are still diverse. Several of these companies are part of a push toward customization in healthcare, an area of biotech set for growth for companies with strong portfolios and experienced leadership.

Here are several startups we considered:

• Syrona Health: Personalized healthcare for women in the workplace

• Anja Health: Personalized umbilical cord blood banking and stem cell preservation

• Alfie: A weight loss program focused on men’s health that coordinates medical care, coaching, and “community-based competition” to help users lose an average of 15% body weight

• Ankr Health: An artificial intelligence (AI)-enabled telehealth platform that provides personalized side effect education for cancer patients and data collection for their care teams

• Koko — A personalized sleep program to improve at-home sleep analysis and training

• Condition-specific telehealth platforms and programs:

• Reviving Mind: Chronic care management covered by insurance and supporting holistic, community-oriented health care

• Equipt Health: At-home delivery of prescription medical equipment to help manage chronic conditions like obstructive sleep apnea

• LunaJoy: Holistic women’s healthcare management for mental health therapy, counseling, and medication

12 Startups from YC's Winter 2022 Demo Day to Watch

Bobidi: 10x faster AI model improvement

Artificial intelligence (AI) models have become a significant tool for firms to improve how well and rapidly they process data. Bobidi helps AI-reliant firms evaluate their models, boosting data insights in less time and reducing data analysis expenditures. The business has created a gamified community that offers a bug bounty for AI, incentivizing community members to test and find weaknesses in clients' AI models.

Magna: DeFi investment management and token vesting

Magna delivers rapid, secure token vesting so consumers may turn DeFi investments into primitives. Carta for Web3 allows enterprises to effortlessly distribute tokens to staff or investors. The Magna team hopes to allow corporations use locked tokens as collateral for loans, facilitate secondary liquidity so investors can sell shares on a public exchange, and power additional DeFi applications.

Perl Street: Funding for infrastructure

This Fintech firm intends to help hardware entrepreneurs get financing by [democratizing] structured finance, unleashing billions for sustainable infrastructure and next-generation hardware solutions. This network has helped hardware entrepreneurs achieve more than $140 million in finance, helping companies working on energy storage devices, EVs, and creating power infrastructure.

CypherD: Multichain cryptocurrency wallet

CypherD seeks to provide a multichain crypto wallet so general customers can explore Web3 products without knowledge hurdles. The startup's beta app lets consumers access crypto from EVM blockchains. The founders have crypto, financial, and startup experience.

Unravel Carbon: Enterprise carbon tracking and offsetting

Unravel Carbon's AI-powered decarbonization technology tracks companies' carbon emissions. Singapore-based startup focuses on Asia. The software can use any company's financial data to trace the supply chain and calculate carbon tracking, which is used to make regulatory disclosures and suggest carbon offsets.

LunaJoy: Precision mental health for women

LunaJoy helped women obtain mental health support throughout life. The platform combines data science to create a tailored experience, allowing women to access psychotherapy, medication management, genetic testing, and health coaching.

Posh: Automated EV battery recycling

Posh attempts to solve one of the EV industry's largest logistical difficulties. Millions of EV batteries will need to be decommissioned in the next decade, and their precious metals and residual capacity will go unused for some time. Posh offers automated, scalable lithium battery disassembly, making EV battery recycling more viable.

Unai: VR headset with 5x higher resolution

Unai stands apart from metaverse companies. Its VR headgear has five times the resolution of existing options and emphasizes human expression and interaction in a remote world. Maxim Perumal's method of latency reduction powers current VR headsets.

Palitronica: Physical infrastructure cybersecurity

Palitronica blends cutting-edge hardware and software to produce networked electronic systems that support crucial physical and supply chain infrastructure. The startup's objective is to build solutions that defend national security and key infrastructure from cybersecurity threats.

Reality Defender: Deepfake detection

Reality Defender alerts firms to bogus users and changed audio, video, and image files. Reality Deference's API and web app score material in real time to prevent fraud, improve content moderation, and detect deception.

Micro Meat: Infrastructure for the manufacture of cell-cultured meat

MicroMeat promotes sustainable meat production. The company has created technologies to scale up bioreactor-grown meat muscle tissue from animal cells. Their goal is to scale up cultured meat manufacturing so cultivated meat products can be brought to market feasibly and swiftly, boosting worldwide meat consumption.

Fleetzero: Electric cargo ships

This startup's battery technology will make cargo ships more sustainable and profitable. Fleetzero's electric cargo ships have five times larger profit margins than fossil fuel ships. Fleetzeros' founder has marine engineering, ship operations, and enterprise sales and business experience.

Restructuring and layoffs cause career reevaluation. Your career can benefit.

Once you become institutionalized, the corporate ladder is all you know.

You're bubbled. Extremists term it the corporate Matrix. I'm not so severe because the business world brainwashed me, too.

This boosted my corporate career.

Until I hit bottom.

15 months later, I view my corporate life differently. You may wish to advance professionally. Read this before you do.

Your happiness in the workplace may be deceptive.

I've been fortunate to spend time with corporate aces.

Working for 2.5 years in banking social media gave me some of these experiences. Earlier in my career, I recorded interviews with business leaders.

These people have titles like Chief General Manager and Head Of. New titles brought life-changing salaries.

They seemed happy.

I’d pass them in the hallway and they’d smile or shake my hand. I dreamt of having their life.

The ominous pattern

Unfiltered talks with some of them revealed a different world.

They acted well. They were skilled at smiling and saying the correct things. All had the same dark pattern, though.

Something felt off.

I found my conversations with them were generally for their benefit. They hoped my online antics as a writer/coach would shed light on their dilemma.

They'd tell me they wanted more. When you're one position away from CEO, it's hard not to wonder if this next move will matter.

What really displeased corporate ladder chasers

Before ascending further, consider these.

Zero autonomy

As you rise in a company, your days get busier.

Many people and initiatives need supervision. Everyone expects you to know business details. Weak when you don't. A poor leader is fired during the next restructuring and left to pursue their corporate ambition.

Full calendars leave no time for reflection. You can't have a coffee with a friend or waste a day.

You’re always on call. It’s a roll call kinda life.

Unable to express oneself freely

My 8 years of LinkedIn writing helped me meet these leaders.

I didn't think they'd care. Mistake.

Corporate leaders envied me because they wanted to talk freely again without corporate comms or a PR firm directing them what to say.

They couldn't share their flaws or inspiring experiences.

They wanted to.

Every day they were muzzled eroded by their business dream.

Limited family time

Top leaders had families.

They've climbed the corporate ladder. Nothing excellent happens overnight.

Corporate dreamers rarely saw their families.

Late meetings, customer functions, expos, training, leadership days, team days, town halls, and product demos regularly occurred after work.

Or they had to travel interstate or internationally for work events. They used bags and motel showers.

Initially, they said business class flights and hotels were nice. They'd get bored. 5-star hotels become monotonous.

No hotel beats home.

One leader said he hadn't seen his daughter much. They used to Facetime, but now that he's been gone so long, she rarely wants to talk to him.

So they iPad-parented.

You're miserable without your family.

Held captive by other job titles

Going up the business ladder seems like a battle.

Leaders compete for business gains and corporate advancement.

I saw shocking filthy tricks. Leaders would lie to seem nice.

Captives included top officials.

A different section every week. If they ran technology, the Head of Sales would argue their CRM cost millions. Or an Operations chief would battle a product team over support requests.

After one conflict, another began.

Corporate echelons are antagonistic. Huge pay and bonuses guarantee bad behavior.

Overly centered on revenue

As you rise, revenue becomes more prevalent. Most days, you'd believe revenue was everything. Here’s the problem…

Numbers drain us.

Unless you're a closet math nerd, contemplating and talking about numbers drains your creativity.

Revenue will never substitute impact.

Incapable of taking risks

Corporate success requires taking fewer risks.

Risks can cause dismissal. Risks can interrupt business. Keep things moving so you may keep getting paid your enormous salary and bonus.

Restructuring or layoffs are inevitable. All corporate climbers experience it.

On this fateful day, a small few realize the game they’ve been trapped in and escape. Most return to play for a new company, but it takes time.

Addiction keeps them trapped. You know nothing else. The rest is strange.

You start to think “I’m getting old” or “it’s nearly retirement.” So you settle yet again for the trappings of the corporate ladder game to nowhere.

Should you climb the corporate ladder?

Let me end on a surprising note.

Young people should ascend the corporate ladder. It teaches you business skills and helps support your side gig and (potential) online business.

Don't get trapped, shackled, or muzzled.

Your ideas and creativity become stifled after too much gaming play.

Corporate success won't bring happiness.

Find fulfilling employment that matters. That's it.

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.

Who is AMTD Digital?

An unknown Hong Kong corporation joined the global megacaps worth over $500 billion on Tuesday.

The American Depository Share (ADS) with the ticker code HKD gapped at the open, soaring 25% over the previous closing price as trading began, before hitting an intraday high of $2,555.

At its peak, its market cap was almost $450 billion, more than Facebook parent Meta or Alibaba.

Yahoo Finance reported a daily volume of 350,500 shares, the lowest since the ADS began trading and much below the average of 1.2 million.

Despite losing a fifth of its value on Wednesday, it's still worth more than Toyota, Nike, McDonald's, or Walt Disney.

The company sold 16 million shares at $7.80 each in mid-July, giving it a $1 billion market valuation.

Why the boom?

That market cap seems unjustified.

According to SEC reports, its income-generating assets barely topped $400 million in March. Fortune's emails and calls went unanswered.

Website discloses little about company model. Its one-minute business presentation film uses a Star Wars–like design to sell the company as a "one-stop digital solutions platform in Asia"

The SEC prospectus explains.

AMTD Digital sells a "SpiderNet Ecosystems Solutions" kind of club membership that connects enterprises. This is the bulk of its $25 million annual revenue in April 2021.

Pretax profits have been higher than top line over the past three years due to fair value accounting gains on Appier, DayDayCook, WeDoctor, and five Asian fintechs.

AMTD Group, the company's parent, specializes in investment banking, hotel services, luxury education, and media and entertainment. AMTD IDEA, a $14 billion subsidiary, is also traded on the NYSE.

“Significant volatility”

Why AMTD Digital listed in the U.S. is unknown, as it informed investors in its share offering prospectus that could delist under SEC guidelines.

Beijing's red tape prevents the Sarbanes-Oxley Board from inspecting its Chinese auditor.

This frustrates Chinese stock investors. If the U.S. and China can't achieve a deal, 261 Chinese companies worth $1.3 trillion might be delisted.

Calvin Choi left UBS to become AMTD Group's CEO.

His capitalist background and status as a Young Global Leader with the World Economic Forum don't stop him from praising China's Communist party or celebrating the "glory and dream of the Great Rejuvenation of the Chinese nation" a century after its creation.

Despite having an executive vice chairman with a record of battling corruption and ties to Carrie Lam, Beijing's previous proconsul in Hong Kong, Choi is apparently being targeted for a two-year industry ban by the city's securities regulator after an investor accused Choi of malfeasance.

Some CMIG-funded initiatives produced money, but he didn't give us the proceeds, a corporate official told China's Caixin in October 2020. We don't know if he misappropriated or lost some money.

A seismic anomaly

In fundamental analysis, where companies are valued based on future cash flows, AMTD Digital's mind-boggling market cap is a statistical aberration that should occur once every hundred years.

AMTD Digital doesn't know why it's so valuable. In a thank-you letter to new shareholders, it said it was confused by the stock's performance.

Since its IPO, the company has seen significant ADS price volatility and active trading volume, it said Tuesday. "To our knowledge, there have been no important circumstances, events, or other matters since the IPO date."

Permabears awoke after the jump. Jim Chanos asked if "we're all going to ignore the $400 billion meme stock in the room," while Nate Anderson called AMTD Group "sketchy."

It happened the same day SEC Chair Gary Gensler praised the 20th anniversary of the Sarbanes-Oxley Act, aimed to restore trust in America's financial markets after the Enron and WorldCom accounting fraud scandals.

The run-up revived unpleasant memories of Robinhood's decision to limit retail investors' ability to buy GameStop, regarded as a measure to protect hedge funds invested in the meme company.

Why wasn't HKD's buy button removed? Because retail wasn't behind it?" tweeted Gensler on Tuesday. "Real stock fraud. "You're worthless."