2025 Computer Science Predictions

Summary

In this video, the speaker makes predictions about the future of computer science in 2025, touching on several key topics. They discuss the rise of RISC-V with Linux distributions, the future of cryptography in response to quantum computing, and the growing importance of memory-safe languages like Rust and Go. They predict that AI-generated content, particularly in areas like background music, will become more widespread, while new technologies like AV1 for streaming and LLMs in gaming will emerge. Additionally, they explore the impact of supply chain attacks and the increased adoption of software bill of materials (SBOMs). The video offers an insightful look into the technological advancements shaping the future.

Takeaways

• 😀 Humans are notoriously bad at making predictions due to hindsight bias, and it's crucial to make forward-looking, falsifiable predictions to test prediction abilities.
• 😀 The demand for RISC-V architecture is expected to grow, with at least three major Linux distributions likely to add official support for RISC-V by 2025.
• 😀 Despite RISC-V's growth, the likelihood of the Raspberry Pi Foundation releasing a primary SBC with a RISC-V chip is less than 20%, due to existing partnerships and supply chain concerns.
• 😀 There's a 70% chance that NIST will advise federal agencies and contractors to adopt quantum-resistant cryptographic algorithms by 2025, to prepare for the threat of quantum computing.
• 😀 Although there is a push for memory safety, most existing codebases will remain in memory-unsafe languages like C++ in 2025, but new projects will favor memory-safe languages like Rust or Go.
• 😀 Rust's popularity as a hyped-up programming language may decline by 2025, with Zig potentially seeing greater growth in terms of new users and community engagement.
• 😀 By 2025, at least two major streaming platforms are expected to adopt AV1 as their default codec for HD and 4K content to save bandwidth while maintaining quality.
• 😀 The use of AI upscaling technology will increase, with the potential for AI-assisted video upscaling in both streaming services and content creation tools, including for anime and low-quality videos.
• 😀 AI-generated content, particularly in stock music, will rise in popularity to help creators avoid copyright issues and provide more customized music for background use in content creation.
• 😀 Supply chain attacks, like SolarWinds and Log4J, will continue to be a significant concern, and by 2025, at least three Fortune 500 tech companies will require software bill of materials (SBOMs) as a deliverable for their third-party vendors.

Q & A

Why is it important to make predictions about the future of computer science?

Making predictions helps test our ability to anticipate trends and challenges, as long as the predictions are falsifiable. It encourages forward-thinking and enables us to measure our ability to foresee future outcomes, rather than merely reflecting on hindsight bias.

What is the prediction regarding the growth of Risk 5 within Linux distributions?

The prediction is that at least three major Linux distributions will add official Risk 5 releases, addressing the current lack of packages compiled specifically for Risk 5 and the limited software availability for it.

Why is there a less than 20% chance that the Raspberry Pi Foundation will release a primary SBC running a Risk 5 CPU?

The Raspberry Pi Foundation has a strong relationship with Broadcom, which manufactures ARM CPUs, not Risk 5 CPUs. Additionally, the limitations in software compatibility and the potential supply chain difficulties related to adopting a new architecture make it unlikely for them to transition to Risk 5.

What is the prediction regarding NIST's role in quantum-resistant cryptography?

It is predicted that by 2025, NIST will release new guidelines recommending the use of quantum-resistant cryptographic algorithms, especially for asymmetric encryption like RSA, to prepare for the potential threats posed by quantum computing in the future.

How will memory safety impact programming languages by 2025?

The prediction is that while existing codebases will continue to be in memory-unsafe languages like C++, new codebases will increasingly be written in memory-safe languages like Rust or Go, as the demand for safer and more secure software increases.

What does the speaker think about the future of the programming language Rust?

While Rust has been highly hyped in recent years, it is predicted that by 2025, Rust will see slower growth compared to Zig, another emerging language. This is due to Rust approaching the peak of its hype cycle, with similar patterns seen in other languages like Ruby and Scala.

How will AI-driven technology affect video streaming by 2025?

By 2025, it is predicted that at least two major streaming platforms will adopt AV1 as their default codec for both HD and 4K video. This is due to its efficiency in saving bandwidth while maintaining video quality, making it a logical progression after H.264.

What is the prediction about AI-generated music in content creation?

It is predicted that by 2025, major stock music providers will offer AI-generated copyright-free background music to address copyright challenges, making it easier for content creators to use music without worrying about licensing issues.

What role will LLMs (Large Language Models) play in video games by 2025?

LLMs are expected to enhance NPC (Non-Player Character) interactions in video games by 2025, providing more personalized and dynamic conversations. This could also indirectly impact digital rights management, as the cloud-based models could prevent pirated games from accessing these enhanced interactions.

What is the prediction regarding the adoption of software bill of materials (SBOM) by Fortune 500 companies?

By 2025, it is predicted that at least three major Fortune 500 tech companies will require SBOMs from their third-party vendors. This will help identify vulnerabilities in software dependencies and respond faster to supply chain attacks like SolarWinds and Log4J.