Hello, Human!

Here are the hot topics of the day:

📰 Breaking news: You’ve Noticed. ChatGPT is Becoming Sluggish, Less Inventive, Less Motivated. What’s Behind the Sudden Change in Attitude?

🌩️ A great reading for every nerd: Mastering DALL-E 2. A Complete Guide to Writing Precise Prompts

🔧 AI Tools kit: Plug-in Wizardry. Three Plug-Ins for GPT-4 That Will Make the Chatbot Smarter

🎨 AI exhibit: Have Fun Making Mistakes

Reading time: 2 minutes 37 seconds

📰 Breaking News

You’ve Noticed. ChatGPT is Becoming Sluggish, Less Inventive, Less Motivated. What’s Behind the Sudden Change in Attitude?

The TL:DR of the issue

GPT-4 hasn’t been the upgrade tekkies were expecting. In fact, many users complain it made the chatbot worse.

According to some experts, we’re dealing with a completely different chatbot from the previous version. That’s how much it’s been changed, piece by piece.

• Large Language Models like ChatGPT and Bard become better with scale. They require a lot of data and the power to process that data. The entire process (both the training and the operations) demand exorbitant amounts of energy.

• At the moment, the creators of the biggest invention since fire are LOSING money every time you prompt. In his Senate hearing in June, Sam Altman was saying they would prefer people use ChatGPT less, because it’s expensive to run.

• To mitigate the issue, OpenAI rebuilt ChatGPT into a Mixture of Experts (MoA). According to leaks, they sliced the chatbot into 16 components, each trained in one field. When queried, it only fires up portions instead of the whole system.

• There might be other concerns that got OpenAI to limit their chatbot. The company is already navigating a legal minefield. Developers need to curate the system so that it’s less vulnerable to jailbreaks, refrains from giving medical advice and so on. One of the scenarios is this makes the model less creative.

Read more… 

🌩️ Insight

Mastering DALL-E 2. A Complete Guide to Writing Precise Prompts

AIs are having a hard time generating images that are true to brief. You can end up spending up to 100 dollars trying to get the precise image that you want. That’s where prompt engineering comes in.

So far, I’ve personally never got every single aspect of a prompt reflected in the generated image. here is some amount of miscommunication with the AI that feels unavoidable. The community is just now figuring out how the AI wants to be briefed.

This being said, you don’t need to start from scratch. Some of the stuff has already been figured out.

Read the full article…

⚙️ AI Tools Kit

Plug-in Wizardry. Three Plug-Ins for GPT-4 That Will Make GPT-4 Smarter

If you’re using ChatGPT-4, you will see a drop-down menu for plugins. They’re a one-click install and they change the way GPT-4 responds to you.

Wolfram. Everything Scientific, Plugged Into ChatGPT

Wolfram on its own started as a platform that helps you solve math and science problems.

It combined forces with GPT-4 and it now gives you access to all kinds of real-time data. Wolfram is powerful. It allows you to solve technical problems (math, programming, physics, chemistry, engineering and co), plots graphs for you and has a world of technical knowledge. It gets GPT to make more difficult computation, solve scientific problems with better precision and plot graphs like no other.

ScholarAI, For Reliable Research

Search through 40M+ peer-reviewed papers, explore scientific PDFs and get a new dimension of knowledge. It will make you a list of the most relevant topics for your research, from trusted sources.

AIPRM For Better Prompts on a Range of Tasks

AIPRM is a broad-spectrum prompt management tool to help you get the best results without fine-tuning the prompt. You can select from a list of goals: write the perfect MidJourney prompt, rewrite articles, SEO the helluva text, create a course. ChatGPT will function inside those parameters, without you needing to clarify.

📸 AI Exhibition

Flaws are good

🧭 Join our community

Are you one of the Great Disrupted? We built a support group for people to exchange ideas and learnto thrive in the age of AI.

We will turn this into a great resource for professionals to get assistance and grow!

It was great talking to you!

At the end of the day, you are a like-minded early adopter of a mind-blowing tech. We hope to give you the edge you deserve to ace this fourth industrial revolution!

If this was useful to you, consider recommending us forward to friends and colleagues. We promise to make it worth their while!

The exhibition, from left to right: James Dean, El Che, Noam Chomsky, Ladi Diana, Ozzie, Assange.