Good morning everyone! We're looking at the buzz around ChatGPT, with some users suggesting it's losing its edge - but is it really? In other exciting news, GitHub's Copilot Chat, the new AI-powered coding assistant, is finally out, promising to make developers' lives easier. Lastly, we've got the lowdown on Java 21's new features set to redefine the way we code. We are also switching to 3x a week (Monday, Wednesday, Friday). So, grab your coffee and let's get started on this exciting journey through AI and software tech updates!

- Forrest Knight & Meghanadh Vasireddy

Is ChatGPT Really Getting Dumber?

Is ChatGPT Getting Dumber? That's the question on everyone's mind lately. ChatGPT is an AI language model developed by OpenAI that can generate human-like responses to text prompts. However, recent reports suggest that ChatGPT's performance may be declining.

• Usage Drops: ChatGPT's popularity is declining, and website traffic and app downloads are down. This suggests that users may be losing interest in the platform.

• Complaints from Users: Many users have reported that ChatGPT's responses are becoming less coherent and more repetitive. Some have even suggested that the AI model is "getting lazier".

• Research Findings: A recent study by Stanford University researchers found that ChatGPT's performance has indeed declined over time. The researchers analyzed the model's behavior and concluded that it is "getting stupider."

However, OpenAI, the company behind ChatGPT, has disputed these claims. In a statement to MakeUseOf, OpenAI said that the model's performance has remained consistent over time. They also noted that ChatGPT is just one of many AI language models, and that there is always room for improvement.

So, is ChatGPT really getting dumber? The answer is not clear-cut. While some users and researchers have reported a decline in performance, OpenAI maintains that the model is still performing well. It's possible that the decline in usage and complaints from users are due to other factors, such as increased competition from other AI language models.

GitHub Copilot Chat Beta Now Available

GitHub Copilot is an AI-powered code assistant that suggests lines or entire functions based on the code developers are currently working on. With the introduction of Copilot Chat, developers can now interact with Copilot using natural language, making it easier to obtain the code they need without manual searching.

Here are some key highlights of the GitHub Copilot Chat beta:

• Context-Aware Conversational Assistant: Copilot Chat transforms GitHub Copilot into a context-aware conversational assistant, providing developers with assistance directly within the IDE.

• Efficient Task Execution: Developers can perform tasks such as code searching, code generation, and bug fixing without having to switch between different tools or search for solutions online.

• Natural Language Interaction: Copilot Chat allows developers to interact with Copilot using natural language, making it easier to communicate their coding needs.

• Time and Effort Savings: By leveraging Copilot Chat, developers can save time and effort by quickly obtaining the code they need without manual searching or leaving the editor.

GitHub Copilot has already made a significant impact on the software development community, with the AI tool suggesting code to developers and helping them write code faster. The addition of Copilot Chat takes this assistance to the next level, providing developers with a conversational coding experience within their preferred IDE.

Java 21: Unveiling Exciting New Features and Improvements

In a highly-anticipated move, Oracle is preparing to roll out the Java Development Kit (JDK) 21 in September 2023. This will be the next long-term support release of Oracle's standard Java implementation. Originally intended to include 16 features, the release now brings forward 15, following the decision last month to remove the experimental Shenandoah garbage collector.

Important Features:

• String templates: This new feature allows developers to use placeholders in string literals, which can then be replaced by variable values.

• Structured concurrency: A new programming paradigm aimed at making concurrent code easier to write and manage. It helps ensure that concurrent tasks are started and stopped in a safe and predictable way.

• Virtual threads: Lightweight threads that are designed to improve the performance of I/O-bound applications. They are more efficient and scalable than traditional threads.

• Key Encapsulation Mechanism (KEM) API: A cryptographic API for securely exchanging keys between two parties, designed to be more secure than traditional key exchange methods.

• Record patterns and pattern matching for switch: These features from Project Amber increase the productivity of developers by making code easier to read and write.

• Foreign function and memory API: Derived from Project Panama, these APIs make it possible to interact with native code from the JVM, enhancing interoperability.

• Vector API: Another outcome of Project Panama, this enables the expression of vector computations that compile at runtime to optimal hardware instructions on supported CPU architectures.

• Scoped values: This feature from Project Loom provides a way to define values that are visible to a specific section of code.

Government Regulation on AI

Artificial intelligence (AI) is a rapidly growing field with the potential to revolutionize many aspects of society. However, there are also concerns about the potential harms that AI could cause, such as job displacement, bias, and privacy violations. As a result, there have been calls for government regulation of AI to ensure that it is developed and used in a responsible and ethical manner.

OpenAI's Sam Altman Urges AI Regulation in Senate: Altman acknowledged the potential risks of AI, but also emphasized its benefits. He argued that ensuring the safety of AI is vital to the work of OpenAI and other companies in the field. However, it is unclear how lawmakers will respond to the call for regulation, given the track record of Congress on tech regulations.

Secretary of Defense Remarks for DOD Artificial Intelligence Symposium and Expo: In September 2020, the Secretary of Defense gave a speech on AI, in which he emphasized the need to promote the adoption of AI in a manner consistent with the values shared with allies and partners. This suggests that the government is interested in regulating AI not just for domestic purposes, but also for international cooperation.

Blueprint for an AI Bill of Rights: In October 2022, the White House Office of Science and Technology Policy released a Blueprint for an AI Bill of Rights, which identified five principles that should guide the design, use, and deployment of automated systems to protect the American public in the age of AI. These principles include transparency, accountability, and fairness. The Blueprint is a guide for a society that protects all people from the threats of AI and uses technologies in ways that reinforce our highest values.

"Isaac Newton to AI" Remarks before the National Press Club: In July 2023, the Chair of the Securities and Exchange Commission, Gary Gensler, gave a speech on AI, in which he drew parallels between the development of AI and the scientific discoveries of Isaac Newton. Gensler emphasized the need for government regulation of AI to ensure that it is developed and used in a responsible and ethical manner. He also highlighted the potential risks of AI, such as systemic risk to financial markets and the potential for AI to exacerbate inequality.

Artificial Intelligence, Free Speech, and the First Amendment: There are also concerns about the potential impact of AI on free speech. While content created by generative AI likely has free speech protections, the government can still restrict any AI speech that is traditionally not protected by the First Amendment, such as libel, criminal solicitation, false advertising, and speech that leads to imminent lawless action.

Youtube Spotlight

Coding Challenge Announcement: Tiny Chess Bots

YouTuber Sebastian Lague introduced a coding challenge where participants will modify a C# project to create tiny chess bots. The challenge involves improving the bot's gameplay within a size limitation and submitting their creations for a tournament to find the best and most creative bots.

• Participants are given a C# project with a chess board to play against a computer.

• The task is to modify the "Think" function in the "MyBot.cs" script to improve the computer's moves.

• The challenge has a size limit of 1024 tokens for the bot's code.

• The deadline for submissions is the first of October, and the winner will be determined through a tournament.

More Notes

1. Computer Science’s Vardi wins four prestigious scientific awards

2. Twitter has officially changed its logo to ‘X’ 

3. TikTok launches text-only posts to rival Elon Musk's Twitter

4. U.S. will be short 67,000 chip workers by 2030

5. Microsoft reveals Copilot pricing at Microsoft Inspire: The feature will cost $30 per user per month, on top of normal Microsoft 365 pricing

Was this forwarded to you? Sign Up Here