Japan's startup challenge: 7 reasons

Every day, another Silicon Valley business is bought for a billion dollars, making its founders rich while growing the economy and improving consumers' lives.

Google, Amazon, Twitter, and Medium dominate our daily lives. Tesla automobiles and Moderna Covid vaccinations.

The startup movement started in Silicon Valley, California, but the rest of the world is catching up. Global startup buzz is rising. Except Japan.

644 of CB Insights' 1170 unicorns—successful firms valued at over $1 billion—are US-based. China follows with 302 and India third with 108.

Japan? 6!

1% of US startups succeed. The third-largest economy is tied with small Switzerland for startup success.

Mexico (8), Indonesia (12), and Brazil (12) have more successful startups than Japan (16). South Korea has 16. Yikes! Problem?

Why Don't Startups Exist in Japan More?

Not about money. Japanese firms invest in startups. To invest in startups, big Japanese firms create Silicon Valley offices instead of Tokyo.

Startups aren't the issue either. Local governments are competing to be Japan's Shirikon Tani, providing entrepreneurs financing, office space, and founder visas.

Startup accelerators like Plug and Play in Tokyo, Osaka, and Kyoto, the Startup Hub in Kobe, and Google for Startups are many.

Most of the companies I've encountered in Japan are either local offices of foreign firms aiming to expand into the Japanese market or small businesses offering local services rather than disrupting a staid industry with new ideas.

There must be a reason Japan can develop world-beating giant corporations like Toyota, Nintendo, Shiseido, and Suntory but not inventive startups.

Culture, obviously. Japanese culture excels in teamwork, craftsmanship, and quality, but it hates moving fast, making mistakes, and breaking things.

If you have a brilliant idea in Silicon Valley, quit your job, get money from friends and family, and build a prototype. To fund the business, you approach angel investors and VCs.

Most non-startup folks don't aware that venture capitalists don't want good, profitable enterprises. That's wonderful if you're developing a solid small business to consult, open shops, or make a specialty product. However, you must pay for it or borrow money. Venture capitalists want moon rockets. Silicon Valley is big or bust. Almost 90% will explode and crash. The few successes are remarkable enough to make up for the failures.

Silicon Valley's high-risk, high-reward attitude contrasts with Japan's incrementalism. Japan makes the best automobiles and cleanrooms, but it fails to produce new items that grow the economy.

Changeable? Absolutely. But, what makes huge manufacturing enterprises successful and what makes Japan a safe and comfortable place to live are inextricably connected with the lack of startups.

Barriers to Startup Development in Japan

These are the 7 biggest obstacles to Japanese startup success.

1. Unresponsive Employment Market

While the lifelong employment system in Japan is evolving, the average employee stays at their firm for 12 years (15 years for men at large organizations) compared to 4.3 years in the US. Seniority, not experience or aptitude, determines career routes, making it tough to quit a job to join a startup and then return to corporate work if it fails.

1. Conservative Buyers

Even if your product is buggy and undocumented, US customers will migrate to a cheaper, superior one. Japanese corporations demand perfection from their trusted suppliers and keep with them forever. Startups need income fast, yet product evaluation takes forever.

1. Failure intolerance

Japanese business failures harm lives. Failed forever. It hinders risk-taking. Silicon Valley embraces failure. Build another startup if your first fails. Build a third if that fails. Every setback is viewed as a learning opportunity for success.

4. No Corporate Purchases

Silicon Valley industrial giants will buy fast-growing startups for a lot of money. Many huge firms have stopped developing new goods and instead buy startups after the product is validated.

Japanese companies prefer in-house product development over startup acquisitions. No acquisitions mean no startup investment and no investor reward.

Startup investments can also be monetized through stock market listings. Public stock listings in Japan are risky because the Nikkei was stagnant for 35 years while the S&P rose 14x.

5. Social Unity Above Wealth

In Silicon Valley, everyone wants to be rich. That creates a competitive environment where everyone wants to succeed, but it also promotes fraud and societal problems.

Japan values communal harmony above individual success. Wealthy folks and overachievers are avoided. In Japan, renegades are nearly impossible.

6. Rote Learning Education System

Japanese high school graduates outperform most Americans. Nonetheless, Japanese education is known for its rote memorization. The American system, which fails too many kids, emphasizes creativity to create new products.

1. Immigration.

Immigrants start 55% of successful Silicon Valley firms. Some come for university, some to escape poverty and war, and some are recruited by Silicon Valley startups and stay to start their own.

Japan is difficult for immigrants to start a business due to language barriers, visa restrictions, and social isolation.

How Japan Can Promote Innovation

Patchwork solutions to deep-rooted cultural issues will not work. If customers don't buy things, immigration visas won't aid startups. Startups must have a chance of being acquired for a huge sum to attract investors. If risky startups fail, employees won't join.

Will Japan never have a startup culture?

Once a consensus is reached, Japan changes rapidly. A dwindling population and standard of living may lead to such consensus.

Toyota and Sony were firms with renowned founders who used technology to transform the world. Repeatable.

Silicon Valley is flawed too. Many people struggle due to wealth disparities, job churn and layoffs, and the tremendous ups and downs of the economy caused by stock market fluctuations.

The founders of the 10% successful startups are heroes. The 90% that fail and return to good-paying jobs with benefits are never mentioned.

Silicon Valley startup culture and Japanese corporate culture are opposites. Each have pros and cons. Big Japanese corporations make the most reliable, dependable, high-quality products yet move too slowly. That's good for creating cars, not social networking apps.

Can innovation and success be encouraged without eroding social cohesion? That can motivate software firms to move fast and break things while recognizing the beauty and precision of expert craftsmen? A hybrid culture where Japan can make the world's best and most original items. Hopefully.

From oil to attention.

Oil has been the most important commodity for a century. It's sparked wars. Pearl Harbor was a preemptive strike to guarantee Japanese access to Indonesian oil, and it made desert tribes rich. Oil's heyday is over. From oil to attention.

We talked about an information economy. In an age of abundant information, what's scarce? Attention. Scale of the world's largest enterprises, wealth of its richest people, and power of governments all stem from attention extraction, monetization, and custody.

Attention-grabbing isn't new. Humans have competed for attention and turned content into wealth since Aeschylus' Oresteia. The internal combustion engine, industrial revolutions in mechanization and plastics, and the emergence of a mobile Western lifestyle boosted oil. Digitization has put wells in pockets, on automobile dashboards, and on kitchen counters, drilling for attention.

The most valuable firms are attention-seeking enterprises, not oil companies. Big Tech dominates the top 4. Tech and media firms are the sheikhs and wildcatters who capture our attention. Blood will flow as the oil economy rises.

Attention to Detail

More than IT and media companies compete for attention. Podcasting is a high-growth, low-barrier-to-entry chance for newbies to gain attention and (for around 1%) make money. Conferences are good for capturing in-person attention. Salesforce paid $30 billion for Slack's dominance of workplace attention, while Spotify is transforming music listening attention into a media platform.

Conferences, newsletters, and even music streaming are artisan projects. Even 130,000-person Comic Con barely registers on the attention economy's Richter scale. Big players have hundreds of millions of monthly users.

Supermajors

Even titans can be disrupted in the attention economy. TikTok is fracking king Chesapeake Energy, a rule-breaking insurgent with revolutionary extraction technologies. Attention must be extracted, processed, and monetized. Innovators disrupt the attention economy value chain.

Attention pre-digital Entrepreneurs commercialized intriguing or amusing stuff like a newspaper or TV show through subscriptions and ads. Digital storage and distribution's limitless capacity drove the initial wave of innovation. Netflix became dominant by releasing old sitcoms and movies. More ad-free content gained attention. By 2016, Netflix was greater than cable TV. Linear scale, few network effects.

Social media introduced two breakthroughs. First, users produced and paid for content. Netflix's economics are dwarfed by TikTok and YouTube, where customers create the content drill rigs that the platforms monetize.

Next, social media businesses expanded content possibilities. Twitter, Facebook, and Reddit offer traditional content, but they transform user comments into more valuable (addictive) emotional content. By emotional resonance, I mean they satisfy a craving for acceptance or anger us. Attention and emotion are mined from comments/replies, piss-fights, and fast-brigaded craziness. Exxon has turned exhaust into heroin. Should we be so linked without a commensurate presence? You wouldn't say this in person. Anonymity allows fraudulent accounts and undesirable actors, which platforms accept to profit from more pollution.

FrackTok

A new entrepreneur emerged as ad-driven social media anger contaminated the water table. TikTok is remaking the attention economy. Short-form video platform relies on user-generated content, although delivery is narrower and less social.

Netflix grew on endless options. Choice requires cognitive effort. TikTok is the least demanding platform since TV. App video plays when opened. Every video can be skipped with a swipe. An algorithm watches how long you watch, what you finish, and whether you like or follow to create a unique streaming network. You can follow creators and respond, but the app is passive. TikTok's attention economy recombination makes it apex predator. The app has more users than Facebook and Instagram combined. Among teens, it's overtaking the passive king, TV.

Externalities

Now we understand fossil fuel externalities. A carbon-based economy has harmed the world. Fracking brought large riches and rebalanced the oil economy, but at a cost: flammable water, earthquakes, and chemical leaks.

TikTok has various concerns associated with algorithmically generated content and platforms. A Wall Street Journal analysis discovered new accounts listed as belonging to 13- to 15-year-olds would swerve into rabbitholes of sex- and drug-related films in mere days. TikTok has a unique externality: Chinese Communist Party ties. Our last two presidents realized the relationship's perils. Concerned about platform's propaganda potential.

No evidence suggests the CCP manipulated information to harm American interests. A headjack implanted on America's youth, who spend more time on TikTok than any other network, connects them to a neural network that may be modified by the CCP. If the product and ownership can't be separated, the app should be banned. Putting restrictions near media increases problems. We should have a reciprocal approach with China regarding media firms. Ban TikTok

It was a conference theme. I anticipated Axel Springer CEO Mathias Döpfner to say, "We're watching them." (That's CEO protocol.) TikTok should be outlawed in every democracy as an espionage tool. Rumored regulations could lead to a ban, and FCC Commissioner Brendan Carr pushes for app store prohibitions. Why not restrict Chinese propaganda? Some disagree: Several renowned tech writers argued my TikTok diatribe last week distracted us from privacy and data reform. The situation isn't zero-sum. I've warned about Facebook and other tech platforms for years. Chewing gum while walking is possible.

The Future

Is TikTok the attention-economy titans' final evolution? The attention economy acts like it. No original content. CNN+ was unplugged, Netflix is losing members and has lost 70% of its market cap, and households are canceling cable and streaming subscriptions in historic numbers. Snap Originals closed in August after YouTube Originals in January.

Everyone is outTik-ing the Tok. Netflix debuted Fast Laughs, Instagram Reels, YouTube Shorts, Snap Spotlight, Roku The Buzz, Pinterest Watch, and Twitter is developing a TikTok-like product. I think they should call it Vine. Just a thought.

Meta's internal documents show that users spend less time on Instagram Reels than TikTok. Reels engagement is dropping, possibly because a third of the videos were generated elsewhere (usually TikTok, complete with watermark). Meta has tried to downrank these videos, but they persist. Users reject product modifications. Kim Kardashian and Kylie Jenner posted a meme urging Meta to Make Instagram Instagram Again, resulting in 312,000 signatures. Mark won't hear the petition. Meta is the fastest follower in social (see Oculus and legless hellscape fever nightmares). Meta's stock is at a five-year low, giving those who opposed my demands to break it up a compelling argument.

Blue Pill

TikTok's short-term dominance in attention extraction won't be stopped by anyone who doesn't hear Hail to the Chief every time they come in. Will TikTok still be a supermajor in five years? If not, YouTube will likely rule and protect Kings Landing.

56% of Americans regularly watch YouTube. Compared to Facebook and TikTok, 95% of teens use Instagram. YouTube users upload more than 500 hours of video per minute, a number that's likely higher today. Last year, the platform garnered $29 billion in advertising income, equivalent to Netflix's total.

Business and biology both value diversity. Oil can be found in the desert, under the sea, or in the Arctic. Each area requires a specific ability. Refiners turn crude into gas, lubricants, and aspirin. YouTube's variety is unmatched. One-second videos to 12-hour movies. Others are studio-produced. (My Bill Maher appearance was edited for YouTube.)

You can dispute in the comment section or just stream videos. YouTube is used for home improvement, makeup advice, music videos, product reviews, etc. You can load endless videos on a topic or creator, subscribe to your favorites, or let the suggestion algo take over. YouTube relies on user content, but it doesn't wait passively. Strategic partners advise 12,000 creators. According to a senior director, if a YouTube star doesn’t post once week, their manager is “likely to know why.”

YouTube's kevlar is its middle, especially for creators. Like TikTok, users can start with low-production vlogs and selfie videos. As your following expands, so does the scope of your production, bringing longer videos, broadcast-quality camera teams and performers, and increasing prices. MrBeast, a YouTuber, is an example. MrBeast made gaming videos and YouTube drama comments.

Donaldson's YouTube subscriber base rose. MrBeast invests earnings to develop impressive productions. His most popular video was a $3.5 million Squid Game reenactment (the cost of an episode of Mad Men). 300 million people watched. TikTok's attention-grabbing tech is too limiting for this type of material. Now, Donaldson is focusing on offline energy with a burger restaurant and cloud kitchen enterprise.

Steps to Take

Rapid wealth growth has externalities. There is no free lunch. OK, maybe caffeine. The externalities are opaque, and the parties best suited to handle them early are incentivized to construct weapons of mass distraction to postpone and obfuscate while achieving economic security for themselves and their families. The longer an externality runs unchecked, the more damage it causes and the more it costs to fix. Vanessa Pappas, TikTok's COO, didn't shine before congressional hearings. Her comms team over-consulted her and said ByteDance had no headquarters because it's scattered. Being full of garbage simply promotes further anger against the company and the awkward bond it's built between the CCP and a rising generation of American citizens.

This shouldn't distract us from the (still existent) harm American platforms pose to our privacy, teenagers' mental health, and civic dialogue. Leaders of American media outlets don't suffer from immorality but amorality, indifference, and dissonance. Money rain blurs eyesight.

Autocratic governments that undermine America's standing and way of life are immoral. The CCP has and will continue to use all its assets to harm U.S. interests domestically and abroad. TikTok should be spun to Western investors or treated the way China treats American platforms: kicked out.

So rich,

Old power paradigm blocks new planetary paradigm

The difference between our reality and the media's reality is like a tale of two worlds. The greatest and worst of times, really.

Expanding information demands complex skills and understanding to separate important information from ignorance and crap. And that's just the start of determining the source's aim.

Trust who? We see people trust liars in public and then be destroyed by their decisions. Mistakes may be devastating.

Many give up and don't trust anyone. Reality is a choice, though. Same risks.

We must separate our needs and wants from reality. Needs and wants have rules. Greed and selfishness create an unlivable planet.

Culturally, we know this, but we ignore it as foolish. Selfish and greedy people obtain what they want, while others suffer.

We invade, plunder, rape, and burn. We establish civilizations by institutionalizing an exploitable underclass and denying its existence. These cultural lies promote greed and selfishness despite their destructiveness.

Controlling parts of society institutionalize these lies as fact. Many of each age are willing to gamble on greed because they were taught to see greed and selfishness as principles justified by prosperity.

Our cultural understanding recognizes the long-term benefits of collaboration and sharing. This older understanding generates an increasing tension between greedy people and those who see its planetary effects.

Survival requires distinguishing between global and regional realities. Simple, yet many can't do it. This is the first time human greed has had a global impact.

In the past, conflict stories focused on regional winners and losers. Losers lose, winners win, etc. Powerful people see potential decades of nuclear devastation as local, overblown, and not personally dangerous.

Mutually Assured Destruction (MAD) was a human choice that required people to acquiesce to irrational devastation. This prevented nuclear destruction. Most would refuse.

A dangerous “solution” relies on nuclear trigger-pullers not acting irrationally. Since then, we've collected case studies of sane people performing crazy things in experiments. We've been lucky, but the climate apocalypse could be different.

Climate disaster requires only continuing current behavior. These actions already cause global harm, but that's not a threat. These activities must be viewed differently.

Once grasped, denying planetary facts is hard to accept. Deniers can't think beyond regional power. Seeing planet-scale is unusual.

Decades of indoctrination defining any planetary perspective as un-American implies communal planetary assets are for plundering. The old paradigm limits any other view.

In the same way, the new paradigm sees the old regional power paradigm as a threat to planetary civilization and lifeforms. Insane!

While MAD relied on leaders not acting stupidly to trigger a nuclear holocaust, the delayed climatic holocaust needs correcting centuries of lunacy. We must stop allowing craziness in global leadership.

Nothing in our acknowledged past provides a paradigm for such. Only primitive people have failed to reach our level of sophistication.

Before European colonization, certain North American cultures built sophisticated regional nations but abandoned them owing to authoritarian cruelty and destruction. They were overrun by societies that saw no wrong in perpetual exploitation. David Graeber's The Dawn of Everything is an example of historical rediscovery, which is now crucial.

From the new paradigm's perspective, the old paradigm is irrational, yet it's too easy to see those in it as ignorant or malicious, if not both. These people are both, but the collapsing paradigm they promote is older or more ingrained than we think.

We can't shift that paradigm's view of a dead world. We must eliminate this mindset from our nations' leadership. No other way will preserve the earth.

Change is occurring. As always with tremendous transition, younger people are building the new paradigm.

The old paradigm's disintegration is insane. The ability to detect errors and abandon their sources is more important than age. This is gaining recognition.

The breakdown of the previous paradigm is not due to senile leadership, but to systemic problems that the current, conservative leadership cannot recognize.

Stop following the old paradigm.

Technical books aren't like novels. We need a new approach to technical texts. I've spent years looking for a decent reading method. I tried numerous ways before finding one that worked. This post explains how I read technical books efficiently.

What Do I Mean When I Say Effective?

Effectiveness depends on the book. Effective implies I know where to find answers after reading a reference book. Effective implies I learned the book's knowledge after reading it.

I use reference books as tools in my toolkit. I won't carry all my tools; I'll merely need them. Non-reference books teach me techniques. I never have to make an effort to use them since I always have them.

Reference books I like:

• Design Patterns: Elements of Reusable Object-Oriented Software

• Refactoring: Improving the Design of Existing Code 
  You can also check My Top Takeaways from Refactoring here.

• Practical Object-Oriented Design in Ruby: An Agile Primer

Non-reference books I like:

• The Pragmatic Programmer: your journey to mastery

• Clean Code: A Handbook of Agile Software Craftsmans

The Approach

Technical books might be overwhelming to read in one sitting. Especially when you have no idea what is coming next as you read. When you don't know how deep the rabbit hole goes, you feel lost as you read. This is my years-long method for overcoming this difficulty.

Whether you follow the step-by-step guide or not, remember these:

• Understand the terminology. Make sure you get the meaning of any terms you come across more than once. The likelihood that a term will be significant increases as you encounter it more frequently.

• Know when to stop. I've always believed that in order to truly comprehend something, I must delve as deeply as possible into it. That, however, is not usually very effective. There are moments when you have to draw the line and start putting theory into practice (if applicable).

• Look over your notes. When reading technical books or documents, taking notes is a crucial habit to develop. Additionally, you must regularly examine your notes if you want to get the most out of them. This will assist you in internalizing the lessons you acquired from the book. And you'll see that the urge to review reduces with time.

Let's talk about how I read a technical book step by step.

0. Read the Foreword/Preface

These sections are crucial in technical books. They answer Who should read it, What each chapter discusses, and sometimes How to Read? This is helpful before reading the book. Who could know the ideal way to read the book better than the author, right?

1. Scanning

I scan the chapter. Fast scanning is needed.

• I review the headings.

• I scan the pictures quickly.

• I assess the chapter's length to determine whether I might divide it into more manageable sections.

2. Skimming

Skimming is faster than reading but slower than scanning.

• I focus more on the captions and subtitles for the photographs.

• I read each paragraph's opening and closing sentences.

• I examined the code samples.

• I attempt to grasp each section's basic points without getting bogged down in the specifics.

• Throughout the entire reading period, I make an effort to make mental notes of what may require additional attention and what may not. Because I don't want to spend time taking physical notes, kindly notice that I am using the term "mental" here. It is much simpler to recall. You may think that this is more significant than typing or writing “Pay attention to X.”

• I move on quickly. This is something I considered crucial because, when trying to skim, it is simple to start reading the entire thing.

3. Complete reading

Previous steps pay off.

• I finished reading the chapter.

• I concentrate on the passages that I mentally underlined when skimming.

• I put the book away and make my own notes. It is typically more difficult than it seems for me. But it's important to speak in your own words. You must choose the right words to adequately summarize what you have read. How do those words make you feel? Additionally, you must be able to summarize your notes while you are taking them. Sometimes as I'm writing my notes, I realize I have no words to convey what I'm thinking or, even worse, I start to doubt what I'm writing down. This is a good indication that I haven't internalized that idea thoroughly enough.

• I jot my inquiries down. Normally, I read on while compiling my questions in the hopes that I will learn the answers as I read. I'll explore those issues more if I wasn't able to find the answers to my inquiries while reading the book.

Bonus!

Best part: If you take lovely notes like I do, you can publish them as a blog post with a few tweaks.

Conclusion

This is my learning journey. I wanted to show you. This post may help someone with a similar learning style. You can alter the principles above for any technical material.

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'm gadget-obsessed. So I shared my top 6 desk gadgets.

These gadgets improve my workflow and are handy for working from home.

Without further ado...

Computer light bar Xiaomi Mi

I've previously recommended the Xiaomi Mi Light Bar, and I still do. It's stylish and convenient.

The Mi bar is a monitor-mounted desk lamp. The lamp's hue and brightness can be changed with a stylish wireless remote.

Changeable hue and brightness make it ideal for late-night work.

Desk Mat 2.

I wasn't planning to include a desk surface in this article, but I find it improves computer use.

The mouse feels smoother and is a better palm rest than wood or glass.

I'm currently using the overkill Razer Goliathus Extended Chroma RGB Gaming Surface, but I like RGB.

Using a desk surface or mat makes computer use more comfortable, and it's not expensive.

Third, the Logitech MX Master 3 Mouse

The Logitech MX Master 3 or any from the MX Master series is my favorite mouse.

The side scroll wheel on these mice is a feature I've never seen on another mouse.

Side scroll wheels are great for spreadsheets and video editing. It would be hard for me to switch from my Logitech MX Master 3 to another mouse. Only gaming is off-limits.

Google Nest 4.

Without a smart assistant, my desk is useless. I'm currently using the second-generation Google Nest Hub, but I've also used the Amazon Echo Dot, Echo Spot, and Apple HomePod Mini.

As a Pixel 6 Pro user, the Nest Hub works best with my phone.

My Nest Hub plays news, music, and calendar events. It also lets me control lights and switches with my smartphone. It plays YouTube videos.

Google Pixel Stand, No. 5

A wireless charger on my desk is convenient for charging my phone and other devices while I work. My desk has two wireless chargers. I have a Satechi aluminum fast charger and a second-generation Google Pixel Stand.

If I need to charge my phone and earbuds simultaneously, I use two wireless chargers. Satechi chargers are well-made and fast. Micro-USB is my only complaint.

The Pixel Stand converts compatible devices into a smart display for adjusting charging speeds and controlling other smart devices. My Pixel 6 Pro charges quickly. Here's my video review.

6. Anker Power Bank

Anker's 65W charger is my final recommendation. This online find was a must-have. This can charge my laptop and several non-wireless devices, perfect for any techie!

The charger has two USB-A ports and two USB-C ports, one with 45W and the other with 20W, so it can charge my iPad Pro and Pixel 6 Pro simultaneously.

Summary

These are some of my favorite office gadgets. My kit page has an updated list.

Links to the products mentioned in this article are in the appropriate sections. These are affiliate links.

You're up! Share the one desk gadget you can't live without and why.