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.

Apple's search engine has long been rumored. Recent Google developments may confirm the rumor. Is Apple about to become Google's biggest rival?

Here's a video:

People noted Apple's changes in 2020. AppleBot, a web crawler that downloads and caches Internet content, was more active than in the last five years.

Apple hired search engine developers, including ex-Googlers, such as John Giannandrea, Google's former search chief.

Apple also changed the way iPhones search. With iOS 14, Apple's search results arrived before Google's.

These facts fueled rumors that Apple was developing a search engine.

Apple and Google Have a Contract

Many skeptics said Apple couldn't compete with Google. This didn't affect the company's competitiveness.

Apple is the only business with the resources and scale to be a Google rival, with 1.8 billion active devices and a $2 trillion market cap.

Still, people doubted that due to a license deal. Google pays Apple $8 to $12 billion annually to be the default iPhone and iPad search engine.

Apple can't build an independent search product under this arrangement.

Why would Apple enter search if it's being paid to stay out?

Ironically, this partnership has many people believing Apple is getting into search.

A New Default Search Engine May Be Needed

Google was sued for antitrust in 2020. It is accused of anticompetitive and exclusionary behavior. Justice wants to end Google's monopoly.

Authorities could restrict Apple and Google's licensing deal due to its likely effect on market competitiveness. Hence Apple needs a new default search engine.

Apple Already Has a Search Engine

The company already has a search engine, Spotlight.

Since 2004, Spotlight has aired. It was developed to help users find photos, documents, apps, music, and system preferences.

Apple's search engine could do more than organize files, texts, and apps.

Spotlight Search was updated in 2014 with iOS 8. Web, App Store, and iTunes searches became available. You could find nearby places, movie showtimes, and news.

This search engine has subsequently been updated and improved. Spotlight added rich search results last year.

If you search for a TV show, movie, or song, photos and carousels will appear at the top of the page.

This resembles Google's rich search results.

When Will the Apple Search Engine Be Available?

When will Apple's search launch? Robert Scoble says it's near.

Scoble tweeted a number of hints before this year's Worldwide Developer Conference.

Scoble bases his prediction on insider information and deductive reasoning. January 2023 is expected.

Will you use Apple's search engine?

Las Vegas hosted CES 2023. This third and last look at CES 2023 previews upcoming consumer electronics trends that will be crucial for market share.

Definitely start with ICT. Qualcomm CEO Cristiano Amon spoke to CNBC from Las Vegas on China's crackdown and the company's automated driving systems for electric vehicles (EV). The business showed a concept car and its latest Snapdragon processor designs, which offer expanded digital interactions through SalesForce-partnered CRM platforms.

Qualcomm CEO Meets SK Hynix Vice Chairman at CES 2023 On Jan. 6, SK hynix Inc.'s vice chairman and co-CEO Park Jung-ho discussed strengthening www.businesskorea.co.kr.

Electrification is reviving Michigan's automobile industry. Michigan Local News reports that $14 billion in EV and battery manufacturing investments will benefit the state. The report also revealed that the Strategic Outreach and Attraction Reserve (SOAR) fund had generated roughly $1 billion for the state's automotive sector.

Michigan to "dominate" EV battery manufacturing after $2B investment. Michigan spent $2 billion to safeguard www.mlive.com.

Ars Technica is great for technology, society, and the future. After CES 2023, Jonathan M. Gitlin published How many electric car chargers are enough? Read about EV charging network issues and infrastructure spending. Politics aside, rapid technological advances enable EV charging network expansion in American cities and abroad.

New research says US needs 8x more EV chargers by 2030. Electric vehicle skepticism—which is widespread—is fundamentally about infrastructure. arstechnica.com

Finally, the UNEP's The Future of Electric Vehicles and Material Resources: A Foresight Brief. Understanding how lithium-ion batteries will affect EV sales is crucial. Climate change affects EVs in various ways, but electrification and mining trends stand out because more EVs demand more energy-intensive metals and rare earths. Areas & Producers has been publishing my electrification and mining trends articles. Follow me if you wish to write for the publication.

Producers This magazine analyzes medium.com-related corporate, legal, and international news to examine a paradigm shift.

The Weekend Brief (TWB) will routinely cover tech, industrials, and global commodities in global markets, including stock markets. Read more about the future of key areas and critical producers of the global economy in Areas & Producers.

TotalEnergies, Stellantis Form Automotive Cells Company (ACC) A joint-venture to design and build electric vehicles (EVs) was formed in 2020.

Successful workers don't talk this way.

"I know it's unprofessional, but I can't stop." — Author Sandy Hall

Do you realize your unprofessionalism? Do you care? Self-awareness?

Everyone can improve their unprofessionalism. Some workplace phrases and words shouldn't be said.

People often say out loud what they're thinking. They show insecurity, incompetence, and disrespect.

"Think before you speak," goes the saying.

Some of these phrases are "okay" in certain situations, but you'll lose colleagues' respect if you use them often.

Your word choice. Your tone. Your intentions. They matter.

Choose your words carefully to build work relationships and earn peer respect. You should build positive relationships with coworkers and clients.

These 10 phrases are unprofessional. 

1. That Meeting Really Sucked

Wow! Were you there? You should be responsible if you attended. You can influence every conversation.

Alternatives

Improve the meeting instead of complaining afterward. Make it more meaningful and productive.

2. Not Sure if You Saw My Last Email

Referencing a previous email irritates people. Email follow-up can be difficult. Most people get tons of emails a day, so it may have been buried, forgotten, or low priority.

Alternatives

It's okay to follow up, but be direct, short, and let the recipient "save face"

3. Any Phrase About Sex, Politics, and Religion

Discussing sex, politics, and religion at work is foolish. If you discuss these topics, you could face harassment lawsuits.

Alternatives

Keep quiet about these contentious issues. Don't touch them.

4. I Know What I’m Talking About

Adding this won't persuade others. Research, facts, and topic mastery are key to persuasion. If you're knowledgeable, you don't need to say this.

Alternatives

Please don’t say it at all. Justify your knowledge.

5. Per Our Conversation

This phrase sounds like legal language. You seem to be documenting something legally. Cold, stern, and distant. "As discussed" sounds inauthentic.

Alternatives

It was great talking with you earlier; here's what I said.

6. Curse-Word Phrases

Swearing at work is unprofessional. You never know who's listening, so be careful. A child may be at work or on a Zoom or Teams call. Workplace cursing is unacceptable.

Alternatives

Avoid adult-only words.

7. I Hope This Email Finds You Well

This is a unique way to wish someone well. This phrase isn't as sincere as the traditional one. When you talk about the email, you're impersonal.

Alternatives

Genuinely care for others.

8. I Am Really Stressed

Happy, strong, stress-managing coworkers are valued. Manage your own stress. Exercise, sleep, and eat better.

Alternatives

Everyone has stress, so manage it. Don't talk about your stress.

9. I Have Too Much to Do

You seem incompetent. People think you can't say "no" or have poor time management. If you use this phrase, you're telling others you may need to change careers.

Alternatives

Don't complain about your workload; just manage it.

10. Bad Closing Salutations

"Warmly," "best," "regards," and "warm wishes" are common email closings. This conclusion sounds impersonal. Why use "warmly" for finance's payment status?

Alternatives

Personalize the closing greeting to the message and recipient. Use "see you tomorrow" or "talk soon" as closings.

Bringing It All Together

These 10 phrases are unprofessional at work. That meeting sucked, not sure if you saw my last email, and sex, politics, and religion phrases.

Also, "I know what I'm talking about" and any curse words. Also, avoid phrases like I hope this email finds you well, I'm stressed, and I have too much to do.

Successful workers communicate positively and foster professionalism. Don't waste chances to build strong work relationships by being unprofessional.

“Unprofessionalism damages the business reputation and tarnishes the trust of society.” — Pearl Zhu, an American author

This post is a summary. Read full article here

Managers spend much time in 1:1s. Most team members meet with supervisors regularly. 1:1s can help create relationships and tackle tough topics. Few appreciate the 1:1 format's potential. Most of the time, that potential is spent on small talk, surface-level updates, and ranting (Ugh, the marketing team isn’t stepping up the way I want them to).

What if you used that time to have deeper conversations and important insights? What if change was easy?

This post introduces a new 1:1 format to help you dive deeper, faster, and develop genuine relationships without losing impact.

A 1:1 is a chat, you would assume. Why use structure to talk to a coworker? Go! I know how to talk to people. I can write. I've always written. Also, This article was edited by Zoe.

Before you discard something, ask yourself if there's a good reason not to try anything new. Is the 1:1 only a talk, or do you want extra benefits? Try the steps below to discover more.

I. Reflection (5 minutes)

Context-free, broad comments waste time and are useless. Instead, give team members 5 minutes to write these 3 prompts.

1. What's effective?

2. What is decent but could be improved?

3. What is broken or missing?

Why these? They encourage people to be honest about all their experiences. Answering these questions helps people realize something isn't working. These prompts let people consider what's working.

Why take notes? Because you get more in less time. Will you feel awkward sitting quietly while your coworker writes? Probably. Persevere. Multi-task. Take a break from your afternoon meeting marathon. Any awkwardness will pay off.

What happens? After a few minutes of light conversation, create a template like the one given here and have team members fill in their replies. You can pre-share the template (with the caveat that this isn’t meant to take much prep time). Do this with your coworker: Answer the prompts. Everyone can benefit from pondering and obtaining guidance.

This step's output.

Part II: Talk (10-20 minutes)

Most individuals can explain what they see but not what's behind an answer. You don't like a meeting. Why not? Marketing partnership is difficult. What makes working with them difficult? I don't recommend slandering coworkers. Consider how your meetings, decisions, and priorities make work harder. The excellent stuff too. You want to know what's humming so you can reproduce the magic.

First, recognize some facts.

• Real power dynamics exist. To encourage individuals to be honest, you must provide a safe environment and extend clear invites. Even then, it may take a few 1:1s for someone to feel secure enough to go there in person. It is part of your responsibility to admit that it is normal.

• Curiosity and self-disclosure are crucial. Most leaders have received training to present themselves as the authorities. However, you will both benefit more from the dialogue if you can be open and honest about your personal experience, ask questions out of real curiosity, and acknowledge the pertinent sacrifices you're making as a leader.

• Honesty without bias is difficult and important. Due to concern for the feelings of others, people frequently hold back. Or if they do point anything out, they do so in a critical manner. The key is to be open and unapologetic about what you observe while not presuming that your viewpoint is correct and that of the other person is incorrect.

Let's go into some prompts (based on genuine conversations):

• “What do you notice across your answers?”

• “What about the way you/we/they do X, Y, or Z is working well?”

• “ Will you say more about item X in ‘What’s not working?’”

• “I’m surprised there isn’t anything about Z. Why is that?”

• “All of us tend to play some role in maintaining certain patterns. How might you/we be playing a role in this pattern persisting?”

• “How might the way we meet, make decisions, or collaborate play a role in what’s currently happening?”

Consider the preceding example. What about the Monday meeting isn't working? Why? or What about the way we work with marketing makes collaboration harder? Remember to share your honest observations!

Third section: observe patterns (10-15 minutes)

Leaders desire to empower their people but don't know how. We also have many preconceptions about what empowerment means to us and how it works. The next phase in this 1:1 format will assist you and your team member comprehend team power and empowerment. This understanding can help you support and shift your team member's behavior, especially where you disagree.

How to? After discussing the stated responses, ask each team member what they can control, influence, and not control. Mark their replies. You can do the same, adding colors where you disagree.

This step's output.

Next, consider the color constellation. Discuss these questions:

• Is one color much more prevalent than the other? Why, if so?

• Are the colors for the "what's working," "what's fine," and "what's not working" categories clearly distinct? Why, if so?

• Do you have any disagreements? If yes, specifically where does your viewpoint differ? What activities do you object to? (Remember, there is no right or wrong in this. Give explicit details and ask questions with curiosity.)

Example: Based on the colors, you can ask, Is the marketing meeting's quality beyond your control? Were our marketing partners consulted? Are there any parts of team decisions we can control? We can't control people, but have we explored another decision-making method? How can we collaborate and generate governance-related information to reduce work, even if the requirement for prep can't be eliminated?

Consider the top one or two topics for this conversation. No 1:1 can cover everything, and that's OK. Focus on the present.

Part IV: Determine the next step (5 minutes)

Last, examine what this conversation means for you and your team member. It's easy to think we know the next moves when we don't.

Like what? You and your teammate answer these questions.

1. What does this signify moving ahead for me? What can I do to change this? Make requests, for instance, and see how people respond before thinking they won't be responsive.

2. What demands do I have on other people or my partners? What should I do first? E.g. Make a suggestion to marketing that we hold a monthly retrospective so we can address problems and exchange input more frequently. Include it on the meeting's agenda for next Monday.

Close the 1:1 by sharing what you noticed about the chat. Observations? Learn anything?

Yourself, you, and the 1:1

As a leader, you either reinforce or disrupt habits. Try this template if you desire greater ownership, empowerment, or creativity. Consider how you affect surrounding dynamics. How can you expect others to try something new in high-stakes scenarios, like meetings with cross-functional partners or senior stakeholders, if you won't? How can you expect deep thought and relationship if you don't encourage it in 1:1s? What pattern could this new format disrupt or reinforce?

Fight reluctance. First attempts won't be ideal, and that's OK. You'll only learn by trying.

Decentralized perpetual may be the next crypto market boom; with tons of perpetual popping up, let's look at two protocols that offer organic, non-inflationary yields.

Decentralized derivatives exchanges' market share has increased tenfold in a year, but it's still 2% of CEXs'. DEXs have a long way to go before they can compete with centralized exchanges in speed, liquidity, user experience, and composability.

I'll cover gains.trade and GMX protocol in Polygon, Avalanche, and Arbitrum. Both protocols support leveraged perpetual crypto, stock, and Forex trading.

Why these protocols?

Decentralized GMX Gains protocol

Organic yield: path to sustainability

I've never trusted Defi's non-organic yields. Example: XYZ protocol. 20–75% of tokens may be set aside as farming rewards to provide liquidity, according to tokenomics.

Say you provide ETH-USDC liquidity. They advertise a 50% APR reward for this pair, 10% from trading fees and 40% from farming rewards. Only 10% is real, the rest is "Ponzi." The "real" reward is in protocol tokens.

Why keep this token? Governance voting or staking rewards are promoted services.

Most liquidity providers expect compensation for unused tokens. Basic psychological principles then? — Profit.

Nobody wants governance tokens. How many out of 100 care about the protocol's direction and will vote?

Staking increases your token's value. Currently, they're mostly non-liquid. If the protocol is compromised, you can't withdraw funds. Most people are sceptical of staking because of this.

"Free tokens," lack of use cases, and skepticism lead to tokens moving south. No farming reward protocols have lasted.

It may have shown strength in a bull market, but what about a bear market?

What is decentralized perpetual?

A perpetual contract is a type of futures contract that doesn't expire. So one can hold a position forever.

You can buy/sell any leveraged instruments (Long-Short) without expiration.

In centralized exchanges like Binance and coinbase, fees and revenue (liquidation) go to the exchanges, not users.

Users can provide liquidity that traders can use to leverage trade, and the revenue goes to liquidity providers.

Gains.trade and GMX protocol are perpetual trading platforms with a non-inflationary organic yield for liquidity providers.

GMX protocol

GMX is an Arbitrum and Avax protocol that rewards in ETH and Avax. GLP uses a fast oracle to borrow the "true price" from other trading venues, unlike a traditional AMM.

GLP and GMX are protocol tokens. GLP is used for leveraged trading, swapping, etc.

GLP is a basket of tokens, including ETH, BTC, AVAX, stablecoins, and UNI, LINK, and Stablecoins.

GLP composition on arbitrum

GLP composition on Avalanche

GLP token rebalances based on usage, providing liquidity without loss.

Protocol "runs" on Staking GLP. Depending on their chain, the protocol will reward users with ETH or AVAX. Current rewards are 22 percent (15.71 percent in ETH and the rest in escrowed GMX) and 21 percent (15.72 percent in AVAX and the rest in escrowed GMX). escGMX and ETH/AVAX percentages fluctuate.

Where is the yield coming from?

Swap fees, perpetual interest, and liquidations generate yield. 70% of fees go to GLP stakers, 30% to GMX. Organic yields aren't paid in inflationary farm tokens.

Escrowed GMX is vested GMX that unlocks in 365 days. To fully unlock GMX, you must farm the Escrowed GMX token for 365 days. That means less selling pressure for the GMX token.

GMX's status

These are the fees in Arbitrum in the past 11 months by GMX.

GMX works like a casino, which increases fees. Most fees come from Margin trading, which means most traders lose money; this money goes to the casino, or GLP stakers.

Strategies

My personal strategy is to DCA into GLP when markets hit bottom and stake it; GLP will be less volatile with extra staking rewards.

GLP YoY return vs. naked buying

Let's say I invested $10,000 in BTC, AVAX, and ETH in January.

• BTC price: 47665$

• ETH price: 3760$

• AVAX price: $145

Current prices

• BTC $21,000 (Down 56 percent )

• ETH $1233 (Down 67.2 percent )

• AVAX $20.36 (Down 85.95 percent )

Your $10,000 investment is now worth around $3,000.

How about GLP? My initial investment is 50% stables and 50% other assets ( Assuming the coverage ratio for stables is 50 percent at that time)

Without GLP staking yield, your value is $6500.

Let's assume the average APR for GLP staking is 23%, or $1500. So 8000$ total. It's 50% safer than holding naked assets in a bear market.

In a bull market, naked assets are preferable to GLP.

Short farming using GLP

Simple GLP short farming.

You use a stable asset as collateral to borrow AVAX. Sell it and buy GLP. Even if GLP rises, it won't rise as fast as AVAX, so we can get yields.

Let's do the maths

You deposit $10,000 USDT in Aave and borrow Avax. Say you borrow $8,000; you sell it, buy GLP, and risk 20%.

After a year, ETH, AVAX, and BTC rise 20%. GLP is $8800. $800 vanishes. 20% yields $1600. You're profitable. Shorting Avax costs $1600. (Assumptions-ETH, AVAX, BTC move the same, GLP yield is 20%. GLP has a 50:50 stablecoin/others ratio. Aave won't liquidate

In naked Avax shorting, Avax falls 20% in a year. You'll make $1600. If you buy GLP and stake it using the sold Avax and BTC, ETH and Avax go down by 20% - your profit is 20%, but with the yield, your total gain is $2400.

Issues with GMX

GMX's historical funding rates are always net positive, so long always pays short. This makes long-term shorts less appealing.

Oracle price discovery isn't enough. This limitation doesn't affect Bitcoin and ETH, but it affects less liquid assets. Traders can buy and sell less liquid assets at a lower price than their actual cost as long as GMX exists.

As users must provide GLP liquidity, adding more assets to GMX will be difficult. Next iteration will have synthetic assets.

Gains Protocol

Best leveraged trading platform. Smart contract-based decentralized protocol. 46 crypto pairs can be leveraged 5–150x and 10 Forex pairs 5–1000x. $10 DAI @ 150x (min collateral x leverage pos size is $1500 DAI). No funding fees, no KYC, trade DAI from your wallet, keep funds.

DAI single-sided staking and the GNS-DAI pool are important parts of Gains trading. GNS-DAI stakers get 90% of trading fees and 100% swap fees. 10 percent of trading fees go to DAI stakers, which is currently 14 percent!

Trade volume

When a trader opens a trade, the leverage and profit are pulled from the DAI pool. If he loses, the protocol yield goes to the stakers.

If the trader's win rate is high and the DAI pool slowly depletes, the GNS token is minted and sold to refill DAI. Trader losses are used to burn GNS tokens. 25%+ of GNS is burned, making it deflationary.

Due to high leverage and volatility of crypto assets, most traders lose money and the protocol always wins, keeping GNS deflationary.

Gains uses a unique decentralized oracle for price feeds, which is better for leverage trading platforms. Let me explain.

Gains uses chainlink price oracles, not its own price feeds. Chainlink oracles only query centralized exchanges for price feeds every minute, which is unsuitable for high-precision trading.

Gains created a custom oracle that queries the eight chainlink nodes for the current price and, on average, for trade confirmation. This model eliminates every-second inquiries, which waste gas but are more efficient than chainlink's per-minute price.

This price oracle helps Gains open and close trades instantly, eliminate scam wicks, etc.

Other benefits include:

• Stop-loss guarantee (open positions updated)

• No scam wicks

• Spot-pricing

• Highest possible leverage

• Fixed-spreads. During high volatility, a broker can increase the spread, which can hit your stop loss without the price moving.

• Trade directly from your wallet and keep your funds.

• >90% loss before liquidation (Some platforms liquidate as little as -50 percent)

• KYC-free

• Directly trade from wallet; keep funds safe

Further improvements

GNS-DAI liquidity providers fear the impermanent loss, so the protocol is migrating to its own liquidity and single staking GNS vaults. This allows users to stake GNS without permanent loss and obtain 90% DAI trading fees by staking. This starts in August.

Their upcoming improvements can be found here.

Gains constantly add new features and change pairs. It's an interesting protocol.

Conclusion

Next bull run, watch decentralized perpetual protocols. Effective tokenomics and non-inflationary yields may attract traders and liquidity providers. But still, there is a long way for them to develop, and I don't see them tackling the centralized exchanges any time soon until they fix their inherent problems and improve fast enough.

Read the full post here.