Fixing problems isn't enough.

Numerous articles and how-to guides on value propositions focus on fixing consumer concerns.

Contrary to popular opinion, addressing customer pain rarely suffices. Win your market category too.

Core Value Statement

Value proposition usually means a product's main value.

Its how your product solves client problems. The product's core.

Answering these questions creates a relevant core value proposition:

• What tasks is your customer trying to complete? (Jobs for clients)

• How much discomfort do they feel while they perform this? (pains)

• What would they like to see improved or changed? (gains)

After that, you create products and services that alleviate those pains and give value to clients.

Value Proposition by Category

Your product belongs to a market category and must follow its regulations, regardless of its value proposition.

Creating a new market category is challenging. Fitting into customers' product perceptions is usually better than trying to change them.

New product users simplify market categories. Products are labeled.

Your product will likely be associated with a collection of products people already use.

Example: IT experts will use your communication and management app.

If your target clients think it's an advanced mail software, they'll compare it to others and expect things like:

• comprehensive calendar

• spam detectors

• adequate storage space

• list of contacts

• etc.

If your target users view your product as a task management app, things change. You can survive without a contact list, but not status management.

Find out what your customers compare your product to and if it fits your value offer. If so, adapt your product plan to dominate this market. If not, try different value propositions and messaging to put the product in the right context.

Finished Value Proposition

A comprehensive value proposition is when your solution addresses user problems and wins its market category.

Addressing simply the primary value proposition may produce a valuable and original product, but it may struggle to cross the chasm into the mainstream market. Meeting expectations is easier than changing views.

Without a unique value proposition, you will drown in the red sea of competition.

To conclude:

1. Find out who your target consumer is and what their demands and problems are.

2. To meet these needs, develop and test a primary value proposition.

3. Speak with your most devoted customers. Recognize the alternatives they use to compare you against and the market segment they place you in.

4. Recognize the requirements and expectations of the market category.

5. To meet or surpass category standards, modify your goods.

Great products solve client problems and win their category.

Learn or earn

In 2020, I started software engineering. My base wage has progressed as follows:

Amazon (2020): $112,000

Microsoft (2021): $123,000

Google (2022): $169,000

I didn't major in math, but those jumps appear more than a 7% wage increase. Here's a deeper look at the three.

The Three Categories of Compensation

Most software engineering compensation packages at IT organizations follow this format.

Minimum Salary

Base salary is pre-tax income. Most organizations give a base pay. This is paid biweekly, twice monthly, or monthly.

Recruiting Bonus

Sign-On incentives are one-time rewards to new hires. Companies need an incentive to switch. If you leave early, you must pay back the whole cost or a pro-rated amount.

Equity

Equity is complex and requires its own post. A company will promise to give you a certain amount of company stock but when you get it depends on your offer. 25% per year for 4 years, then it's gone.

If a company gives you $100,000 and distributes 25% every year for 4 years, expect $25,000 worth of company stock in your stock brokerage on your 1 year work anniversary.

Performance Bonus

Tech offers may include yearly performance bonuses. Depends on performance and funding. I've only seen 0-20%.

Engineers' overall compensation usually includes:

Base Salary + Sign-On + (Total Equity)/4 + Average Performance Bonus

Amazon: (TC: 150k)

Base Pay System

Amazon pays Seattle employees monthly on the first work day. I'd rather have my money sooner than later, even if it saves processing and pay statements.

The company upped its base pay cap from $160,000 to $350,000 to compete with other tech companies.

Performance Bonus

Amazon has no performance bonus, so you can work as little or as much as you like and get paid the same. Amazon is savvy to avoid promising benefits it can't deliver.

Sign-On Bonus

Amazon gives two two-year sign-up bonuses. First-year workers could receive $20,000 and second-year workers $15,000. It's probably to make up for the company's strange equity structure.

If you leave during the first year, you'll owe the entire money and a prorated amount for the second year bonus.

Equity

Most organizations prefer a 25%, 25%, 25%, 25% equity structure. Amazon takes a different approach with end-heavy equity:

• the first year, 5%

• 15% after one year.

• 20% then every six months

We thought it was constructed this way to keep staff longer.

Microsoft (TC: 185k)

Base Pay System

Microsoft paid biweekly.

Gainful Performance

My offer letter suggested a 0%-20% performance bonus. Everyone will be satisfied with a 10% raise at year's end.

But misleading press where the budget for the bonus is doubled can upset some employees because they won't earn double their expected bonus. Still barely 10% for 2022 average.

Sign-On Bonus

Microsoft's sign-on bonus is a one-time payout. The contract can require 2-year employment. You must negotiate 1 year. It's pro-rated, so that's fair.

Equity

Microsoft is one of those companies that has standard 25% equity structure. Except if you’re a new graduate.

In that case it’ll be

• 25% six months later

• 25% each year following that

New grads will acquire equity in 3.5 years, not 4. I'm guessing it's to keep new grads around longer.

Google (TC: 300k)

Base Pay Structure

Google pays biweekly.

Performance Bonus

Google's offer letter specifies a 15% bonus. It's wonderful there's no cap, but I might still get 0%. A little more than Microsoft’s 10% and a lot more than Amazon’s 0%.

Sign-On Bonus

Google gave a 1-year sign-up incentive. If the contract is only 1 year, I can move without any extra obligations.

Not as fantastic as Amazon's sign-up bonuses, but the remainder of the package might compensate.

Equity

We covered Amazon's tail-heavy compensation structure, so Google's front-heavy equity structure may surprise you.

Annual structure breakdown

• 33% Year 1

• 33% Year 2

• 22% Year 3

• 12% Year 4

The goal is to get them to Google and keep them there.

Final Thoughts

This post hopefully helped you understand the 3 firms' compensation arrangements.

There's always more to discuss, such as refreshers, 401k benefits, and business discounts, but I hope this shows a distinction between these 3 firms.

Sketch or Figma?

Designers are pissed.

The beast ate the beauty.

Figma deserves $20B.

Do designers deserve Adobe?

Adobe devours new creative tools and spits them out with a slimy Adobe aftertaste.

• Frame.io — $1.3B

• Magento — $1.7B

• Macromedia — $3.6B

Nothing compares to the risky $20B acquisition.

If they can't be beaten, buy them.

And then make them boring.

Adobe's everywhere.

Like that friend who dabbles in everything creatively, there's not enough time to master one thing.

Figma was Adobe's thigh-mounted battle axe.

• a UX design instrument with a sizable free tier.

• a UX design tool with a simple and quick user interface.

• a tool for fluid collaboration in user experience design.

• a web-based UX design tool that functions well.

• a UX design tool with a singular goal of perfection.

UX design software that replaced Adobe XD.

Adobe XD could do many of Figma's things, but it didn't focus on the details. This is a major issue when working with detail-oriented professionals.

UX designers.

Design enthusiasts first used Figma. More professionals used it. Institutions taught it. Finally, major brands adopted Figma.

Adobe hated that.

Adobe dispatched a team of lawyers to resolve the Figma issue, as big companies do. Figma didn’t bite for months.

Oh no.

Figma resisted.

Figma helped designers leave Adobe. Figma couldn't replace Photoshop, but most designers used it to remove backgrounds.

Online background removal tools improved.

The Figma problem grew into a thorn, a knife, and a battle ax in Adobe's soft inner thigh.

Figma appeared to be going public. Adobe couldn’t allow that. It bought Figma for $20B during the IPO drought.

Adobe has a new issue—investors are upset.

The actual cause of investors' ire toward Adobe

Spoiler: The math just doesn’t add up.

According to Adobe's press release, Figma's annual recurring revenue (ARR) is $400M and growing rapidly.

The $20B valuation requires a 50X revenue multiple, which is unheard of.

Venture capitalists typically use:

• 10% to 29% growth per year: ARR multiplied by 1 to 5

• 30% to 99% growth per year: ARR multiplied by 6 to 10

• 100% to 400% growth per year: ARR multiplied by 10 to 20

Showing an investor a 50x multiple is like telling friends you saw a UFO. They'll think you're crazy.

Adobe's stock fell immediately after the acquisition because it didn't make sense to a number-cruncher.

Designers started a Tweet storm in the digital town hall where VCs and designers often meet.

Adobe acquired Workfront for $1.5 billion at the end of 2020. This purchase made sense for investors.

Many investors missed the fact that Adobe is acquiring Figma not only for its ARR but also for its brilliant collaboration tech.

Adobe could use Figmas web app technology to make more products web-based to compete with Canva.

Figma's high-profile clients could switch to Adobe's enterprise software.

However, questions arise:

• Will Adobe make Figma boring?

• Will Adobe tone down Figma to boost XD?

• Would you ditch Adobe and Figma for Sketch?

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.

I started smoking in 2002 at IIT BHU. Most of us thought it was enjoyable at first. I didn't realize the cost later.

In 2005, during my final semester, I lost my father. Suddenly, I felt more accountable for my mother and myself.

I quit before starting my first job in Bangalore. I didn't see any smoking friends in my hometown for 2 months before moving to Bangalore.

For the next 5-6 years, I had no regimen and smoked only when drinking.

Due to personal concerns, I started smoking again after my 2011 marriage. Now smoking was a constant guilty pleasure.

I smoked 3-4 cigarettes a day, but never in front of my family or on weekends. I used to excuse this with pride! First office ritual: smoking. Even with guilt, I couldn't stop this time because of personal concerns.

After 8-9 years, in mid 2019, a personal development program solved all my problems. I felt complete in myself. After this, I just needed one cigarette each day.

The hardest thing was leaving this final cigarette behind, even though I didn't want it.

James Clear's Atomic Habits was published last year. I'd only read 2-3 non-tech books before reading this one in August 2021. I knew everything but couldn't use it.

In April 2022, I realized the compounding effect of a bad habit thanks to my subconscious mind. 1 cigarette per day (excluding weekends) equals 240 = 24 packs per year, which is a lot. No matter how much I did, it felt negative.

Then I applied the 2nd principle of this book, identifying the trigger. I tried to identify all the major triggers of smoking. I found social drinking is one of them & If I am able to control it during that time, I can easily control it in other situations as well. Going further whenever I drank, I was pre-determined to ignore the craving at any cost. Believe me, it was very hard initially but gradually this craving started fading away even with drinks.

I've been smoke-free for 3 months. Now I know a bad habit's effects. After realizing the power of habits, I'm developing other good habits which I ignored all my life.