Daily AI Newsletter: GitHub can now automatically fix code vulnerabilities

In today’s email:

🧩 Can you beat an AI at WORDLE?
📚 Under The Hood: How OpenAI's Sora Model Works
🤩 8 Google Employees Invented Modern AI. Here’s the Inside Story
🧰 9 new AI-powered tools and resources. Make sure to check the online version for the full list of tools.

GitHub’s latest AI tool can automatically fix code vulnerabilities

GitHub has launched a new beta feature called code-scanning autofix, aimed at identifying and correcting security vulnerabilities in code. This innovative tool integrates GitHub's Copilot real-time capabilities with CodeQL, its semantic code analysis engine, to automatically remedy over two-thirds of detected vulnerabilities in several programming languages, including JavaScript, Typescript, Java, and Python. Available to GitHub Advanced Security customers, this feature is designed to streamline the debugging process, allowing developers to focus on other tasks while enhancing security measures.

GitHub's code-scanning auto-fix utilizes the CodeQL engine, initially released to the public in 2019, to scan for vulnerabilities in code before execution. The tool, which has evolved from its initial version, employs a mix of heuristics and GitHub Copilot APIs to propose fixes, leveraging OpenAI's GPT-4 model to generate solutions and their explanations. While GitHub assures the effectiveness of the auto-fix suggestions, it acknowledges that a small fraction might not accurately address the codebase or the vulnerabilities.

The introduction of code-scanning autofix represents a significant step forward in automating code security and efficiency. By reducing the time developers spend on remediation, GitHub aims to allow development teams to allocate more resources to strategic initiatives, simultaneously enabling security teams to focus on broader protective measures. This new feature underscores GitHub's commitment to advancing development practices and enhancing the security of the coding environment.

News. Without Motives. That’s 1440. Join Today!

Frustrated by biased news? Tired of trying to sift through 100s of news sources to find the facts? Leave behind the opinions and receive a daily digest of pure facts. Say hello to 1440! Delve into politics, business, science, sports, culture, and more. Join the other 3 million Americans who read 1440 each morning.

Learn More

Characters now have a voice

Character AI is excited to introduce Character Voice, a new feature enhancing user interaction by allowing Characters to speak in 1:1 chats. This addition is part of a broader effort to create a multimodal interface, making interactions more intuitive and engaging. Character Voice, initially available to the c.ai+ Community, is now free for all users. The Voice library offers a selection of user-generated and pre-made voices that can be assigned to Characters or chats, and users can also create their own voices through the app.

The Voice feature supports customization and sharing, enabling users to create, search, and apply public Voices to their Characters or use them in personal chats. Currently, the feature supports English, with plans to include more languages to cater to the global user base. Character AI is committed to improving user experience through feedback and continuous updates.

Character AI emphasizes safety and encourages users to visit their Safety Center to learn about the safety measures implemented for Character Voice. The platform aims to empower users to be the protagonists of their own stories, using Character Voice to bring a new dimension to their interactions and creativity. For more details, users are encouraged to check the FAQ section on Character AI's website.

Key Stable Diffusion Researchers Leave Stability AI As Company Flounders

Key researchers behind the Stable Diffusion text-to-image model, including Robin Rombach, Andreas Blattmann, and Dominik Lorenz, have left Stability AI, the company they joined after developing their model at a German university. Their departure marks a significant setback for Stability AI, which has been facing challenges with executive exits and financial strains despite a successful $100 million seed funding in 2022. The researchers' work had played a critical role in the AI boom, contributing significantly to Stability AI's initial success.

The exodus from Stability AI isn't limited to the research team; several top executives and vice presidents have also resigned over the past year. The company's financial struggles are evident, with reports indicating high expenditure rates surpassing revenue, leading to difficulties in managing operational costs, including payroll. Despite these challenges, Stability AI managed to secure a $50 million lifeline from Intel and has been attempting to navigate through its financial and legal hurdles, including copyright infringement lawsuits.

Stability AI's turmoil reflects broader challenges within the AI industry, highlighted by internal conflicts, legal issues, and the competitive nature of the sector. The departure of key personnel from Stability underscores the volatile nature of the tech industry, where rapid changes and intense competition can significantly impact companies, regardless of their initial success or innovative contributions.

Other stuff

8 Google Employees Invented Modern AI. Here’s the Inside Story
Can you beat an AI at WORDLE?
Google Deepmind Introduced: TacticAI, an AI assistant for football tactics
AI Launchpad - Unlock the full potential of your AI business in 6 weeks
Under The Hood: How OpenAI's Sora Model Works
Saudi Arabia is reportedly in talks with VC firms like Andreessen Horowitz to create a mammoth $40 billion AI fund
LLMs: RAG vs. Fine-Tuning
RAFT: A new way to teach LLMs to be better at RAG
Neuralink demonstrating “Telepathy”
Worldsim and jailbreaking Claude 3
Introducing HPT: A Groundbreaking Family of Leading Multimodal LLMs