A fascinating business conversation.

This week at Web Summit, I had mentor hour.

Mentor hour connects startups with experienced entrepreneurs.

The YC-selected founder who mentored me had grown his company to $100 million in 18 months.

I had 45 minutes to question him.

I've compiled this.

Context

Founder's name is Zack.

After working in private equity, Zack opted to acquire an MBA.

Surrounded by entrepreneurs at a prominent school, he decided to become one himself.

Unsure how to proceed, he bet on two horses.

On one side, he received an offer from folks who needed help running their startup owing to lack of time. On the other hand, he had an idea for a SaaS to start himself.

He just needed to validate it.

Validating

Since Zack's proposal helped companies, he contacted university entrepreneurs for comments.

He contacted university founders.

Once he knew he'd correctly identified the problem and that people were willing to pay to address it, he started developing.

He earned $100k in a university entrepreneurship competition.

His plan was evident by then.

The other startup's founders saw his potential and granted him $400k to launch his own SaaS.

Hiring

He started looking for a tech co-founder because he lacked IT skills.

He interviewed dozens and picked the finest.

As he didn't want to wait for his program to be ready, he contacted hundreds of potential clients and got 15 letters of intent promising they'd join up when it was available.

YC accepted him by then.

He had enough positive signals to raise.

Raising

He didn't say how many VCs he called, but he indicated 50 were interested.

He jammed meetings into two weeks to generate pressure and encourage them to invest.

Seed raise: $11 million.

Selling

His objective was to contact as many entrepreneurs as possible to promote his product.

He first contacted startups by scraping CrunchBase data.

Once he had more money, he started targeting companies with ZoomInfo.

His VC urged him not to hire salespeople until he closed 50 clients himself.

He closed 100 and hired a CRO through a headhunter.

Scaling

Three persons started the business.

1. He primarily works in sales.

2. Coding the product was done by his co-founder.

3. Another person performing operational duties.

He regretted recruiting the third co-founder, who was ineffective (could have hired an employee instead).

He wanted his company to be big, so he hired two young marketing people from a competing company.

After validating several marketing channels, he chose PR.

$100 Million and under

He developed a sales team and now employs 30 individuals.

He raised a $100 million Series A.

Additionally, he stated

• He’s been rejected a lot. Like, a lot.

• Two great books to read: Steve Jobs by Isaacson, and Why Startups Fail by Tom Eisenmann.

• The best skill to learn for non-tech founders is “telling stories”, which means sales. A founder’s main job is to convince: co-founders, employees, investors, and customers. Learn code, or learn sales.

Conclusion

I often read about these stories but hardly take them seriously.

Zack was amazing.

Three things about him stand out:

1. His vision. He possessed a certain amount of fire.

2. His vitality. The man had a lot of enthusiasm and spoke quickly and decisively. He takes no chances and pushes the envelope in all he does.

3. His Rolex.

He didn't do all this in 18 months.

Not really.

He couldn't launch his company without private equity experience.

These accounts disregard entrepreneurs' original knowledge.

Hormozi will tell you how he founded Gym Launch, but he won't tell you how he had a gym first, how he worked at uni to pay for his gym, or how he went to the gym and learnt about fitness, which gave him the idea to open his own.

Nobody knows nothing. If you scale quickly, it's probable because you gained information early.

Lincoln said, "Give me six hours to chop down a tree, and I'll spend four sharpening the axe."

Sharper axes cut trees faster.

SaaS spending is 40/40/20. 40% of operating expenses should be R&D, 40% sales and marketing, and 20% G&A. We wanted to see the statistics behind the rules of thumb. Since October 2017, 73 SaaS startups have gone public. Perhaps the rule of thumb should be 30/50/20. The data is below.

30/50/20. R&D accounts for 26% of opex, sales and marketing 48%, and G&A 22%. We think R&D/S&M/G&A should be 30/50/20.

There are outliers. There are exceptions to rules of thumb. Dropbox spent 45% on R&D whereas Zoom spent 13%. Zoom spent 73% on S&M, Dropbox 37%, and Bill.com 28%. Snowflake spent 130% of revenue on S&M, while their EBITDA margin is -192%.

G&A shouldn't stand out. Minimize G&A spending. Priorities should be product development and sales. Cloudflare, Sendgrid, Snowflake, and Palantir spend 36%, 34%, 37%, and 43% on G&A.

Another myth is that COGS is 20% of revenue. Median and averages are 29%.

Where is the profitability? Data-driven operating income calculations were simplified (Revenue COGS R&D S&M G&A). 20 of 73 IPO businesses reported operational income. Median and average operating income margins are -21% and -27%.

As long as you're growing fast, have outstanding retention, and marquee clients, you can burn cash since recurring income that doesn't churn is a valuable annuity.

The data was compelling overall. 30/50/20 is the new 40/40/20 for more established SaaS enterprises, unprofitability is alright as long as your business is expanding, and COGS can be somewhat more than 20% of revenue.

Price Intelligently put out amazing content on pricing your SaaS product. This blog's link to the whole report is worth reading. Our key takeaways are below.

Don't base prices on the competition. Competitor-based pricing has clear drawbacks. Their pricing approach is yours. Your company offers customers something unique. Otherwise, you wouldn't create it. This strategy is static, therefore you can't add value by raising prices without outpricing competitors. Look, but don't touch is the competitor-based moral. You want to know your competitors' prices so you're in the same ballpark, but they shouldn't guide your selections. Competitor-based pricing also drives down prices.

Value-based pricing wins. This is customer-based pricing. Value-based pricing looks outward, not inward or laterally at competitors. Your clients are the best source of pricing information. By valuing customer comments, you're focusing on buyers. They'll decide if your pricing and packaging are right. In addition to asking consumers about cost savings or revenue increases, look at data like number of users, usage per user, etc.

Value-based pricing increases prices. As you learn more about the client and your worth, you'll know when and how much to boost rates. Every 6 months, examine pricing.

Cloning top customers. You clone your consumers by learning as much as you can about them and then reaching out to comparable people or organizations. You can't accomplish this without knowing your customers. Segmenting and reproducing them requires as much detail as feasible. Offer pricing plans and feature packages for 4 personas. The top plan should state Contact Us. Your highest-value customers want more advice and support.

Question your 4 personas. What's the one item you can't live without? Which integrations matter most? Do you do analytics? Is support important or does your company self-solve? What's too cheap? What's too expensive?

Not everyone likes per-user pricing. SaaS organizations often default to per-user analytics. About 80% of companies utilizing per-user pricing should use an alternative value metric because their goods don't give more value with more users, so charging for them doesn't make sense.

At least 3:1 LTV/CAC. Break even on the customer within 2 years, and LTV to CAC is greater than 3:1. Because customer acquisition costs are paid upfront but SaaS revenues accrue over time, SaaS companies face an early financial shortfall while paying back the CAC.

ROI should be >20:1. Indeed. Ensure the customer's ROI is 20x the product's cost. Microsoft Office costs $80 a year, but consumers would pay much more to maintain it.

A/B Testing. A/B testing is guessing. When your pricing page varies based on assumptions, you'll upset customers. You don't have enough customers anyway. A/B testing optimizes landing pages, design decisions, and other site features when you know the problem but not pricing.

Don't discount. It cheapens the product, makes it permanent, and increases churn. By discounting, you're ruining your pricing analysis.

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.

Economic mutual destruction.

Multibillion-dollar corporations never bothered with an independent audit, and they all lied about their balance sheets.

It's easy to forget that the Big Four accounting firms are lousy fraud enablers. Just because they sign off on your books doesn't mean you're not a hoax waiting to erupt.

This is *crazy* Capitalism depends on independent auditors. Rich folks need to know their financial advisers aren't lying. Rich folks usually succeed.

No accounting. EY, KPMG, PWC, and Deloitte make more money consulting firms than signing off on their accounts.

The Big Four sign off on phony books because failing to make friends with unscrupulous corporations may cost them consulting contracts.

The Big Four are the only firms big enough to oversee bankruptcy when they sign off on fraudulent books, as they did for Carillion in 2018. All four profited from Carillion's bankruptcy.

The Big Four are corrupt without any consequences for misconduct. Who can forget when KPMG's top management was fined millions for helping auditors cheat on ethics exams?

Consulting and auditing conflict. Consultants help a firm cover its evil activities, such as tax fraud or wage theft, whereas auditors add clarity to a company's finances. The Big Four make more money from cooking books than from uncooking them, thus they are constantly embroiled in scandals.

If a major scandal breaks, it may bring down the entire sector and substantial parts of the economy. Jim Peterson explains system risk for The Dig.

The Big Four are voluntary private partnerships where accountants invest their time, reputations, and money. If a controversy threatens the business, partners who depart may avoid scandal and financial disaster.

When disaster looms, each partner should bolt for the door, even if a disciplined stay-and-hold posture could weather the storm. This happened to Arthur Andersen during Enron's collapse, and a 2006 EU report recognized the risk to other corporations.

Each partner at a huge firm knows how much dirty laundry they've buried in the company's garden, and they have well-founded suspicions about what other partners have buried, too. When someone digs, everyone runs.

If a firm confronts substantial litigation damages or enforcement penalties, it could trigger the collapse of one of the Big Four. That would be bad news for the firm's clients, who would have trouble finding another big auditor.

Most of the world's auditing capacity is concentrated in four enormous, brittle, opaque, compromised organizations. If one of them goes bankrupt, the other three won't be able to take on its clients.

Peterson: Another collapse would strand many of the world's large public businesses, leaving them unable to obtain audit views for their securities listings and regulatory compliance.

Count Down: The Past, Present, and Uncertain Future of the Big Four Accounting Firms is in its second edition.

https://www.emerald.com/insight/publication/doi/10.1108/9781787147003