⚡️ OpenAI's Preparedness Framework

Good Morning. We got some news. After receiving a bunch of feedback proposing decreased frequency, we’re pivoting Supercharged AI to go out 3x per week instead of daily. Expect upcoming editions to be juicier and signal-dense. Anyway, we have a lot to cover today—let’s dive in.

Today’s Highlights:

• Google Gemini is not even as good as GPT-3.5 Turbo

• AI Decodes Mental Images from Brain Activity

• Stability AI launches subscription plans for commercial use

DEEP DIVE

OpenAI Introduces Framework to Address AI Safety and Risks

Amidst a wave of AI developments, OpenAI, the creators of ChatGPT, have introduced a methodical approach to AI safety: the "Preparedness Framework," to help measure, manage, and mitigate hazards associated with advancing AI technologies.

The release comes during a tumultuous phase for the organization, spotlighting the vital role of governance and ethical considerations as AI systems grow more sophisticated and influential.

Key Highlights of the Framework:

• Usage of dynamic risk scorecards for assessing potential dangers linked with AI models, including those that could facilitate cyberattacks or autonomous weaponry.

• Emphasis on solid, data-based evaluations over speculative risks to focus on reliable predictions and effective measures.

• Continuous refinement ensured through new data, feedback loops, and shared expertise within the AI community.

Striking Differences with Anthropic's Policy:

The introduction of OpenAI's Framework contrasts noticeably with the Responsible Scaling Policy by Anthropic—a competitor and similarly influential AI lab. Whereas Anthropic embeds safety directly into its progression policies, OpenAI opts for a more dynamic and flexible system that depends on triggered reviews.

While OpenAI's approach grants adaptability and discretionary leeway, it could leave more room for human error, a gap that seems narrower under Anthropic's more regulated policy.

Despite these methods varying in application, their existence underscores a broader, integral push toward prioritizing AI safety. It marks a pivotal evolution in the field, shifting some focus from capabilities to the ethical implementation and safeguarding of AI progress.

Moving forward, unified efforts in establishing robust safety standards will be paramount to benefit society and harness the full potential of AI.

Keep reading here.

PUNCHLINES

The AI Crescent: Revo Capital stirs the pot with $100 million, eyeing golden AI opportunities in Turkey's startup scene.

Binary Brewmaster: Japanese brewery using generative AI to dream up new beverages.

New Power Couple: IBM ties the knot with StreamSets and WebMethods in a $2.3 billion deal to boost its hybrid cloud prowess.

Generative Gap: Google Gemini is not even as good as GPT-3.5 Turbo, researchers find.

Ghostwriting 2.0: New York City council member-elect used AI to answer questions.

TLDR

AI Decodes Mental Images from Brain Activity: Japanese researchers successfully use AI to convert brain signals into images, unveiling a brain decoding technology that visualizes mental imagery of natural scenes. The pioneering technique utilizes fMRI data to train a generative AI, with promises of advancing medical and welfare applications.

Microsoft Copilot introduces songwriting with Suno: Microsoft enhances its AI chatbot with Suno's music-making capabilities. Users can type prompts to create full songs, complete with lyrics and vocals. Available through Microsoft Edge, the partnership aims to democratize music creation.

Stability AI launches subscription plans for commercial use: Stability AI unveils a tiered subscription service for its AI models, including Stable Diffusion. While retaining a free option, paid plans cater to small businesses and enterprises, providing commercial usage rights and early access to new AI releases.

AI Detects Autism from Retinal Images with Unmatched Accuracy: Korean researchers develop an AI that diagnoses childhood autism flawlessly using eye pictures, a promising leap for objective ASD screening.

TRENDING TOOLS

🎤 VoiceDual: Transform your voice with groundbreaking AI technology

🔍 Vexa Search: Dive into knowledge exploration with image-driven search capabilities

🧘‍♂️ Mental-: Strengthen men's mental health on-the-go

📈 Smartly: Optimize your marketing campaigns with AI-enhanced workflows

📝 MeetGeek: Capture and distill meeting knowledge automatically with video and transcription AI

That’s all for today—if you have any questions or something interesting to share, please reply to this email. We’d love to hear from you!

P.S. If you want to sign up for the Supercharged newsletter or share it with a friend, you can find us here.