Calmness is a vital life skill.

It aids communication. It boosts creativity and performance.

I've studied calm people's habits for years. Commonalities:

Have mastered the art of self-humor.

Protectors take their job seriously, draining the room's energy.

They are fixated on positive pursuits like making cool things, building a strong physique, and having fun with others rather than on depressing influences like the news and gossip.

Every day, spend at least 20 minutes moving, whether it's walking, yoga, or lifting weights.

Discover ways to take pleasure in life's challenges.

Since perspective is malleable, they change their view.

Set your own needs first.

Stressed people neglect themselves and wonder why they struggle.

Prioritize self-care.

Don't ruin your life to please others.

Make something.

Calm people create more than react.

They love creating beautiful things—paintings, children, relationships, and projects.

Don’t hold their breath.

If you're stressed or angry, you may be surprised how much time you spend holding your breath and tightening your belly.

Release, breathe, and relax to find calm.

Stopped rushing.

Rushing is disadvantageous.

Calm people handle life better.

Are aware of their own dietary requirements.

They avoid junk food and eat foods that keep them healthy, happy, and calm.

Don’t take anything personally.

Stressed people control everything.

Self-conscious.

Calm people put others and their work first.

Keep their surroundings neat.

Maintaining an uplifting and clutter-free environment daily calms the mind.

Minimise negative people.

Calm people are ruthless with their boundaries and avoid negative and drama-prone people.

https://tomconnor.me/wp-content/uploads/2021/08/10x-Engineer-Mental-Models.pdf

I keep returning to the same mental models and tricks after writing and reading about a wide range of topics.

Top 12 mental models

12.

Survival bias - We perceive the surviving population as remarkable, yet they may have gotten there through sheer grit.

Survivorship bias affects us in many situations. Our retirement fund; the unicorn business; the winning team. We often study and imitate the last one standing. This can lead to genuine insights and performance improvements, but it can also lead us astray because the leader may just be lucky.

11.

The Helsinki Bus Theory - How to persevere Buss up!

Always display new work, and always be compared to others. Why? Easy. Keep riding. Stay on the fucking bus.

10.

Until it sticks… Turning up every day… — Artists teach engineers plenty. Quality work over a career comes from showing up every day and starting.

9.

WRAP decision making process (Heath Brothers)

Decision-making WRAP Model:

W — Widen your Options

R — Reality test your assumptions

A — Attain Distance

P — Prepare to be wrong or Right

8.

Systems for knowledge worker excellence - Todd Henry and Cal Newport write about techniques knowledge workers can employ to build a creative rhythm and do better work.

Todd Henry's FRESH framework:

1. Focus: Keep the start in mind as you wrap up.

2. Relationships: close a loop that's open.

3. Pruning is an energy.

4. Set aside time to be inspired by stimuli.

5. Hours: Spend time thinking.

7.

Black Box Thinking…..

BBT is learning from mistakes. Science has transformed the world because it constantly updates its theories in light of failures. Complexity guarantees failure. Do we learn or self-justify?

6.

The OODA Loop - Competitive advantage

O: Observe: collect the data. Figure out exactly where you are, what’s happening.

O: Orient: analyze/synthesize the data to form an accurate picture.

D: Decide: select an action from possible options

A: Action: execute the action, and return to step (1)

Boyd's approach indicates that speed and agility are about information processing, not physical reactions. They form feedback loops. More OODA loops improve speed.

5.

Know your Domain 

Leaders who try to impose order in a complex situation fail; those who set the stage, step back, and allow patterns to develop win.

https://vimeo.com/640941172?embedded=true&source=vimeo_logo&owner=11999906

4.

The Three Critical Gaps

• Information Gap - The discrepancy between what we know and what we would like to know

• Gap in Alignment - What individuals actually do as opposed to what we wish them to do

• Effects Gap - the discrepancy between our expectations and the results of our actions

3.

Theory of Constraints — The Goal  - To maximize system production, maximize bottleneck throughput.

• Goldratt creates a five-step procedure:

1. Determine the restriction

2. Improve the restriction.

3. Everything else should be based on the limitation.

4. Increase the restriction

5. Go back to step 1 Avoid letting inertia become a limitation.

Any non-constraint improvement is an illusion.

2.

Serendipity and the Adjacent Possible - Why do several amazing ideas emerge at once? How can you foster serendipity in your work?

You need specialized abilities to reach to the edge of possibilities, where you can pursue exciting tasks that will change the world. Few people do it since it takes a lot of hard work. You'll stand out if you do.

Most people simply lack the comfort with discomfort required to tackle really hard things. At some point, in other words, there’s no way getting around the necessity to clear your calendar, shut down your phone, and spend several hard days trying to make sense of the damn proof.

1.

Boundaries of failure - Rasmussen's accident model.

Rasmussen modeled this. It has economic, workload, and performance boundaries.

The economic boundary is a company's profit zone. If the lights are on, you're within the economic boundaries, but there's pressure to cut costs and do more.

Performance limit reflects system capacity. Taking shortcuts is a human desire to minimize work. This is often necessary to survive because there's always more labor.

Both push operating points toward acceptable performance. Personal or process safety, or equipment performance.

If you exceed acceptable performance, you'll push back, typically forcefully.

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.

What they'll do is scarier.

My acquaintance recently joined a VC-funded startup. Money, equity, and upside possibilities were nice, but he had a nagging dread.

They just secured a $40M round and are hiring like crazy to prepare for their IPO in two years. All signals pointed to this startup's (a B2B IT business in a stable industry) success, and its equity-holding workers wouldn't pass that up.

Five months after starting the work, my friend struggled with leaving. We might overlook the awful culture and long hours at the proper price. This price plus the company's fate and survival abilities sent my friend departing in an unpleasant unplanned resignation before jumping on yet another sinking ship.

This affects founders. This affects VC-backed companies (and all businesses). This affects anyone starting, buying, or running a business.

Here's the under-the-table approach that's draining VC capital, leaving staff terrified (or jobless), founders rattled, and investors upset. How to recognize, solve, and avoid it

The unsettling reality behind door #1

You can't raise money off just your looks, right? If "looks" means your founding team's expertise, then maybe. In my friend's case, the founding team's strong qualifications and track records won over investors before talking figures.

They're hardly the only startup to raise money without a profitable customer acquisition strategy. Another firm raised money for an expensive sleep product because it's eco-friendly. They were off to the races with a few keywords and key players.

Both companies, along with numerous others, elected to invest on product development first. Company A employed all the tech, then courted half their market (they’re a tech marketplace that connects two parties). Company B spent millions on R&D to create a palatable product, then flooded the world with marketing.

My friend is on Company B's financial team, and he's seen where they've gone wrong. It's terrible.

Company A (tech market): Growing? Not quite. To achieve the ambitious expansion they (and their investors) demand, they've poured much of their little capital into salespeople: Cold-calling commission and salary salesmen. Is it working? Considering attrition and companies' dwindling capital, I don't think so.

Company B (green sleep) has been hiring, digital marketing, and opening new stores like crazy. Growing expenses should result in growing revenues and a favorable return on investment; if you grow too rapidly, you may neglect to check that ROI.

Once Company A cut headcount and Company B declared “going concerned”, my friend realized both startups had the same ailment and didn't recognize it.

I shouldn't have to ask a friend to verify a company's cash reserves and profitability to spot a financial problem. It happened anyhow.

The frightening part isn't that investors were willing to invest millions without product-market fit, CAC, or LTV estimates. That's alarming, but not as scary as the fact that startups aren't understanding the problem until VC rounds have dried up.

When they question consultants if their company will be around in 6 months. It’s a red flag. How will they stretch $20M through a 2-year recession with a $3M/month burn rate and no profitability? Alarms go off.

Who's in danger?

In a word, everyone who raised money without a profitable client acquisition strategy or enough resources to ride out dry spells.

Money mismanagement and poor priorities affect every industry (like sinking all your capital into your product, team, or tech, at the expense of probing what customer acquisition really takes and looks like).

This isn't about tech, real estate, or recession-proof luxury products. Fast, cheap, easy money flows into flashy-looking teams with buzzwords, trending industries, and attractive credentials.

If these companies can't show progress or get a profitable CAC, they can't raise more money. They die if they can't raise more money (or slash headcount and find shoestring budget solutions until they solve the real problem).

The kiss of death (and how to avoid it)

If you're running a startup and think raising VC is the answer, pause and evaluate. Do you need the money now?

I'm not saying VC is terrible or has no role. Founders have used it as a Band-Aid for larger, pervasive problems. Venture cash isn't a crutch for recruiting consumers profitably; it's rocket fuel to get you what and who you need.

Pay-to-play isn't a way to throw money at the wall and hope for a return. Pay-to-play works until you run out of money, and if you haven't mastered client acquisition, your cash will diminish quickly.

How can you avoid this bottomless pit? Tips:

• Understand your burn rate

• Keep an eye on your growth or profitability.

• Analyze each and every marketing channel and initiative.

• Make lucrative customer acquisition strategies and satisfied customers your top two priorities. not brand-new products. not stellar hires. avoid the fundraising rollercoaster to save time. If you succeed in these two tasks, investors will approach you with their thirsty offers rather than the other way around, and your cash reserves won't diminish as a result.

Not as much as your grandfather

My family friend always justified expensive, impractical expenditures by saying it was only monopoly money. In business, startups, and especially with money from investors expecting a return, that's not true.

More founders could understand that there isn't always another round if they viewed VC money as their own limited pool. When the well runs dry, you must refill it or save the day.

Venture financing isn't your grandpa's money. A discerning investor has entrusted you with dry powder in the hope that you'll use it wisely, strategically, and thoughtfully. Use it well.