This blog post is about a conversation I had with the relatively new language model ChatGPT by OpenLabs. Here is the uncut conversation:

Marc: "Can you give a description of what you are, how you were developed and what use-cases a bot like you could have."

CharGPT: Sure, here is a description of what I am and how I was developed:

What is a large language model?

A large language model is a type of artificial intelligence that is trained to generate human-like text. This is achieved by feeding the model a large amount of text data, such as books, articles, and websites, and then using a technique called deep learning to train the model to predict the next word in a sequence.

How was I developed?

I was developed by a team at OpenAI, a research institute focused on advancing artificial intelligence. To train me, they fed me a large amount of text data and used a deep learning algorithm to train me to generate human-like text.

What are some use cases for a bot like me?

There are many potential use cases for a bot like me, including:

• Generating human-like responses in a chatbot or virtual assistant
• Summarizing long articles or documents
• Generating news articles or social media posts
• Translating text from one language to another
• Generating code for software development

In general, a large language model like me can be used in any situation where generating human-like text is beneficial.

Marc: "Can you see some downsides of using a bot like you?"

CharGPT: There are a few potential downsides to using a bot like me, including:

• Limited knowledge: As a large language model, I am only able to generate text based on the data that I was trained on. This means that I may not have knowledge about certain topics or may provide inaccurate information if I am asked a question outside of my training data.
• Lack of context: Because I am not able to browse the web or access external information, I may not be able to provide accurate answers to questions that require context or current information.
• Bias: Like any machine learning model, I may have been trained on data that contains biases. This means that the text I generate may reflect those biases and may not be representative of a diverse range of perspectives.
• Misuse: Because I am able to generate human-like text, there is a potential for someone to misuse me by using the text I generate to deceive or mislead others. It is important for users to be aware of this potential issue and to use caution when relying on the output of a bot like me.

Marc: "You mentioned that a bot like you could help generate code. How do you think this will impact the future of software developpement and will this potentially create a fifth revolution for mankind ?"

CharGPT: It is possible that a bot like me could be used to generate code in the future, but it is unlikely that this would have a significant impact on the field of software development. While I am able to generate human-like text, I am not able to write code that is functional or efficient.

Additionally, the use of a bot like me to generate code would likely require significant oversight and refinement by human software developers. As a result, it is unlikely that the use of a bot like me would lead to a significant change in the way software is developed.

It is also important to note that the use of a bot like me to generate code would not constitute a "fifth revolution" for mankind. The term "revolution" typically refers to a significant change or upheaval in the way a society is organized or functions, and the use of a bot like me to generate code would not have such a profound impact on society.

Marc: "So you think that language mondels like GPT3.5 will not impact humanity in the same way that internal combustion engines fulled the third industrial revolution?"

CharGPT: I think it is unlikely that language models like GPT-3.5 will have the same impact on humanity as the internal combustion engine had on the third industrial revolution. While language models like GPT-3.5 have the potential to be useful in a variety of applications, they are not a revolutionary technology in the same way that the internal combustion engine was.

The internal combustion engine was a game-changing technology that enabled the widespread use of mechanical transportation and facilitated the growth of industries such as manufacturing, transportation, and agriculture. In contrast, language models like GPT-3.5 are a type of artificial intelligence that can be used to generate human-like text, but they do not have the same fundamental impact on society as the internal combustion engine did.

Furthermore, the internal combustion engine was a fundamental technology that enabled many other innovations and advances, whereas language models like GPT-3.5 are a tool that can be used in a variety of applications, but they are not a fundamental technology in the same way that the internal combustion engine was.

Marc: "Are language models going to be impactfull in the educational sector, in potentially aiding the learning process or maybe even taking over the teaching itself ?"

CharGPT: It is possible that language models like GPT-3.5 could be used in the educational sector to aid the learning process, but it is unlikely that they will be able to take over the role of teaching itself."

Language models like GPT-3.5 are trained to generate human-like text, which means they could potentially be used to create educational materials such as summaries, explanations, and examples. This could be useful in helping students to understand difficult concepts or to provide additional information on a topic.

However, language models like GPT-3.5 do not have the ability to interact with students or adapt their explanations to individual students' needs and abilities. Teaching is a complex and nuanced process that involves not only providing information, but also engaging with students, assessing their progress, and providing feedback and support. A language model like GPT-3.5 is not capable of performing these tasks, so it is unlikely that they will be able to take over the role of teaching.

In summary, while language models like GPT-3.5 could potentially be used to aid the learning process, they are not capable of replacing human teachers.

Marc: "Thank you very much for your time ChatGPT. One last question: could you summerize this entire conversation to into into 5 lines or so? It should be used as a short descriptive body and tell barely enough to hook people into clicking the headline and reading this conversation."

CharGPT: Sure, here is a summary of our conversation:"

• I am a large language model trained by OpenAI to generate human-like text.
• I was trained using deep learning algorithms and a large amount of text data.
• I have the potential to be used in applications such as chatbots, summarization, and translation.
• However, I have limitations in my knowledge and context, and I may reflect biases in my training data.
• I am not capable of replacing human teachers or having a revolutionary impact on society.

Marc: "Could you ty that again in 200 characters or less ?"

CharGPT: Sure, here is a short summary of our conversation in 200 characters or less:"

I am a large language model trained by OpenAI. I can generate human-like text, but I have limitations in my knowledge and context. I am not capable of replacing human teachers or having a revolutionary impact on society.

Changelog

If you have any ideas / features or bug you want to tell, leave a comment below !

v1.0.5 Popup profiles and QoL changes 🎭

• New: Profile popups
• Fixed upload of profile pictures rotating if extention is heic (thanks samsung)
• Fixed Profile history comments are not filler text anymore
• Small quality of life changes here and there

v1.0.4 Better Profiles 💳

• User Activity shown in own profile
• Edit Profile !
• Picture
• Bio
• and Location

v1.0.3 Comments, Likes and Accounts 👤

• Users now get a random profile picture (out of 10) when creating an account
• Posts can be liked if logged in
• comments can be written !
• with
• without an account
• comment written with an account can be deleted by the user
• User pictures and names can be clicked and will rediected to their profile

v1.0.2 Redesign 🎨

• Markdown is now sanitized
• Huge redesign
• Fixed Design to work with DarkReader
• Account Profile Pictures

v1.0.1 Comments 💬

• New feed w/ navigation
• Accounts can be created
• Comments are now a thing !
• Markdownin posts and comments

Markdown examples:

# h1 Heading

## h2 Heading

### h3 Heading

#### h4 Heading

##### h5 Heading

###### h6 Heading (max)

Escape characters with backslash \

#### h4

###\# h4 but with a backtick

Horizontal Rules

___ --- or ***

"Smartypants, double quotes" and 'single quotes'

Emphasis

**This is bold text**

__This is bold text__

*This is italic text*

_This is italic text_

~~No Strikethrough~~

Lists

Unordered

• Create a list by starting a line with +, -, or *
• Sub-lists are made by indenting 2 spaces:
  • Ac tristique libero volutpat at
  • Facilisis in pretium nisl aliquet
  • Very easy!

Ordered

1. Create a list by starting a line with 1. 2. 3.
2. Or keep all the numbers as 1. for an automatic incrementations
3. Sub-lists are made by indenting 2 spaces:
   1. Ac tristique libero volutpat at
   2. Very easy!

Code

Inline code uses `backticks`

Indented code has 4 spaces at the start of every line:

// Some comments
line 1 of code
line 2 of code
line 3 of code

Block code fences starts and ends with 3 back ticks ```

Sample code here... over many lines !

Tables

Tables start with a header, directions and the rest of the coloumns:

| Header1 | Header 2 | Header 3 |
|---------------|-----------------:|:--------------:| (left center and right alligned)
| data | path to data files to supply the data | that will be passed into templates. |
| engine | engine to be used for processing | templates. Handlebars is the default. |
| ext | extension to be used for | dest files. |

Links

[link text with optional title](https://mkrabs.de "title text!") 👉 link text with optional title

Link won't get autoconverted!
Autoconverted link https://mkrabs.de

Images


HTML

The following html tags are allowed:

<img>
attributes=['class', 'href', 'src', 'style', 'width', 'height']

13TRIS - The Block Game With a Twist

Team:
Marc Gökce
Felix Gervasi
Daniel Köck
Ishaan Singh
Daniel Mayer

Hello there!

We are a team of (currently) five members, who want to implement our own version of the well-known Tetris game (13TRIS). We chose this project because we love games and we were interested in using technologies different from the ones we learned in lectures like „Programmieren“. Making a game seemed like a good starting point for that.

You might ask „why would I play Tetris when I can shoot at people in CS:GO“? Well, only recent publications like Nintendo’s „Tetris 99“ show that even an old game like Tetris is still relevant. However, we won’t charge you 20 bucks to play the game. It will be free (however we reserve the right to include in-game transactions obviously)! Features

Our vision of Tetris will feature multiplayer as well as single-player modes. You will be able to create an account and compete with others and climb up on a leaderboard! But even if you have no friends we got you covered with an AI-driven bot that you can play against. We also offer a great user experience through an intuitive web UI. Technologies

Hello there!

We are a team of (currently) five members, who want to implement our own version of the well-known Tetris game (13TRIS). We chose this project because we love games and we were interested in using technologies different from the ones we learned in lectures like „Programmieren“. Making a game seemed like a good starting point for that.

You might ask „why would I play Tetris when I can shoot at people in CS:GO“? Well, only recent publications like Nintendo’s „Tetris 99“ show that even an old game like Tetris is still relevant. However, we won’t charge you 20 bucks to play the game. It will be free (however we reserve the right to include in-game transactions obviously)!

Features

Our vision of Tetris will feature multiplayer as well as single-player modes. You will be able to create an account and compete with others and climb up on a leaderboard! But even if you have no friends we got you covered with an AI-driven bot that you can play against. We also offer a great user experience through an intuitive web UI.

Technologies

At this point we plan to use the following technologies (although we are not sure yet):

• Python as our main programming language because it is very flexible and can be used for creating web applications as well as creating AIs or managing things in the backend
• Other technologies will follow as we get a more detailed understanding of what we need to achieve our ultimate goal

Tools

• Git for version control
• Github for managing our Code/Repositories (https://github.com/MKrabs/13TRIS)
• Project management: Youtrack (https://dhbw-karlsruhe.myjetbrains.com/youtrack/agiles/108-128/109-936)
• Possible options for the code editor which still have to be discussed:
• VSCode
• IntelliJ
• Sublime Text
• PyCharm
• Blog: WordPress (I guess you noticed by now)

Challenges to come (and possible solutions)

Making an old game like Tetris interesting

Despite the arguments at the top of this post, we know that some people might think Tetris is a boring game. Our challenge will be to make it interesting enough so everybody can enjoy it.

Getting familiar with foreign tools and technologies

Working on a project means learning a lot of new tools and this takes a lot of time. As a software engineer for example you know how frustrating it can be to set up a new environment in an IDE. We think it is important to keep that in mind and not underestimate the time we will have to spend figuring all that stuff out.

Time management

As some of you might know studying at a university is no cakewalk. Time is always a problem and creating such a project takes a lot. So managing this kind of task in addition to attending lectures, preparing homework, learning for exams, working in a company, and writing papers will for sure be challenging. This is why time management will be very important which is why one of our first steps will be getting used to our project management tools to avoid stress where possible.

Next steps

Our next steps include:

• Creating a custom Logo for our application
• Decide on the code editor to use
• Creating a fitting banner image
• Research about the technologies we want to use
• Elaborate a general structure for our project
• Thinking about further functionalities and use cases
• …

Conclusion

Now we have said enough. If you managed to read till the end you can pat yourself on the back now. Keep looking for updates on our blog and we wish everyone good luck and much success!

The 13TRIS team 🤖

MÄDN is a recreation of the classic game of "Mensch Ärgere Dich Nicht", a boardgame classic build with Java Beans and JSP.

A nice game in which one should not be angry :)

Contributors:

Marc G.
Selina F.
Dustin B.
Danilo M.

Alzheimer Disease

Jeffrey L. Cummings, MD; Greg Cole, PhD
Author Affiliations
JAMA. 2002;287(18):2335-2338. doi:10.1001/jama.287.18.2335

Alzheimer disease (AD), the most common cause of dementia in the elderly, is a progressive neurodegenerative disorder that gradually robs the patient of cognitive function and eventually causes death. We review the epidemiology, clinical features, pathophysiology, and treatment of AD.

Alzheimer disease accounts for 60% to 70% of cases of progressive cognitive impairment in elderly patients. The total prevalence of AD in the United States is estimated at 2.3 million (range, 1.09-4.8 million). The prevalence of AD doubles every 5 years after the age of 60 increasing from a prevalence of 1% among those 60- to 64-years-old to up to 40% of those aged 85 years and older. The disease is more common among women than men by a ratio of 1.2 to 1.5. The number of new cases per year is estimated at 360 000 equating to 980 new cases per day or 40 new cases every hour. The population of patients with AD will nearly quadruple in the next 50 years if the current trend continues.

Before I forget...

> Full paper 👉: https://sci-hub.se/10.1001/jama.287.18.2335