OpenAI GPT-4.5 Rolled Out

In today’s Newsletter,

1. OpenAI ChatGPT 4.5 rolled out

2. ChatGPT Prompts to Creating and Selling High-Converting Affiliate Kits (and Earning $10,000+/Month)

3. AI & TECH NEWS

OpenAI Previews GPT-4.5 (SOURCE)

OpenAI is releasing a "research preview" of GPT-4.5, its newest and largest chat model. This version focuses on expanded knowledge and improved "intuition" through increased unsupervised learning, rather than enhanced reasoning skills.

GPT-4.5 was trained on massive amounts of data using Microsoft Azure AI supercomputers. As a result, it has a broader knowledge base, understands user intent better, and is expected to "hallucinate" (make things up) less often.

Start learning AI in 2025

Keeping up with AI is hard – we get it!

That’s why over 1M professionals read Superhuman AI to stay ahead.

• Get daily AI news, tools, and tutorials

• Learn new AI skills you can use at work in 3 mins a day

• Become 10X more productive

Sign up and start mastering AI

25 ChatGPT Prompts to Creating and Selling High-Converting Affiliate Kits (and Earning $10,000+/Month)

The affiliate marketing world is booming, but standing out requires more than just slapping together some generic promotional materials. A powerful strategy is creating and selling "affiliate kits" – pre-packaged bundles of marketing assets that other marketers can use to promote a specific product or service. These kits save time and effort, making them highly valuable to busy affiliates.

AI & TECH NEWS

Qwen's New QwQ-32B: An Open-Source AI Focused on Advanced Reasoning (Source)

Qwen has launched QwQ-32B, a new open-source AI model designed specifically for strong reasoning skills, particularly in math and coding. This addresses a common weakness in many existing AI systems that struggle with complex, multi-step problems.

QwQ-32B boasts 32.5 billion parameters and advanced features like an extended context length (32,768 tokens), allowing it to handle long, intricate inputs. Crucially, it uses reinforcement learning (RL) during training, focusing on improving performance in math and coding through outcome-based rewards.

This open-source approach contrasts with proprietary models, promoting transparency and community collaboration. Performance data shows QwQ-32B is competitive with state-of-the-art systems in math and code generation, and the RL training helps it avoid common AI pitfalls. QwQ-32B is a significant step for open-source AI, providing a powerful, accessible tool for researchers and developers.

Scientists Skeptical of Google's "AI Co-Scientist" (Source)

Google claims its new Gemini 2.0-based AI tool can act as a "co-scientist," generating research hypotheses. However, many human scientists are doubtful, saying it's unlikely to be useful in real research.

The AI supposedly uses "advanced reasoning" and multiple AI "agents" debating ideas to create research plans. Google touted it as giving scientists "superpowers," but experts point out several problems:

• Scientists say they don't need AI to generate hypotheses; that's the fun part of their job.

• The AI's suggested drug repurposing for leukemia was too vague to be taken seriously by experts.

• Claims of discovering novel treatments for liver fibrosis were false; the suggested drugs were already known.

While the AI might be helpful for summarizing large amounts of research quickly, the concern remains that it might "hallucinate" (make up information). Ultimately, Google's claim of a hypothesis-generating machine, not just a research assistant, is seen as overblown and unwanted by the scientific community.

Google Search Adds Powerful "AI Mode" Powered by Gemini 2.0

Google is enhancing search with "AI Mode," leveraging its Gemini 2.0 AI for smarter results. Key improvements include better AI-generated summaries (AI Overviews) for complex queries, including math, coding, and multimodal (text, image, voice) input. AI overview is available now without requiring to login.

"AI Mode," an experimental feature, handles multi-step questions, comparisons, and conversational follow-ups. It uses real-time data and accepts text, voice, and image input. Google prioritizes accuracy through fact-checking and safety filters.

AI mode is Initially exclusive to Google One AI Premium subscribers, accessible via a dedicated tab/page/app icon.

Please remember

I share actionable guides, AI & tech news, Tools to help you learn AI and also launch a successful side hustle.

Don't miss out!

Mark emails from <OpenAILearning> [email protected] as important and also move this email to primary so that every email land directly in your inbox.

Thanks for your time and consideration,

Shailesh and OpenAILearning Team