Innovation Profs - 7/22/2025

Your weekly guide to generative AI tools and news

Innovation Profs
July 22, 2025

Latest Gen AI News

OpenAI launches a general purpose agent in ChatGPT

Last Thursday OpenAI launched ChatGPT Agent, a new tool that combines the web-navigation features of Operator, the information processing capabilities of Deep Research, and the conversational abilities of ChatGPT to complete complex tasks for users with its own computer. Access to Agent for ChatGPT Plus users is reportedly being rolled out incrementally (we don’t have access to it yet), but reviews of the tool have not been stellar, as one can see from the following headlines:

We’ll keep you posted on its performance when we get our hands on the tool in the coming weeks.

OpenAI Debuts Record Mode for ChatGPT Plus

Regardless of how you may feel about the increased assimilation of AI tools into common productivity workflows, you likely don’t cherish the prospects of taking minutes at your next staff meeting. ChatGPT Plus users can now put that practice behind them with Record Mode, which allows users to record meetings up to 2 hours in length using the ChatGPT desktop app. “After the recording, the data is fed through the company servers, where timestamped references, action items, and summaries are generated.” The audio files are subsequently deleted.

Netflix Uses Generative AI in TV Show for First Time

This week Netflix revealed that it used generative AI to create a VFX sequence of a building collapse in the Argentinian drama “The Eternaut.” According to Netflix co-CEO Ted Sarandos, “[T]hat VFX sequence was completed 10 times faster than it could have been completed with traditional VFX tools and workflows.” While this increase in efficiency is impressive and signals exciting creative possibilities, it also raises uneasy questions about the future of jobs in visual effects.

Quick Hits

Mistral releases Voxtral, its first open source AI audio model | TechCrunch

French startup Mistral has jumped into the audio race with Voxtral, its first open model, aiming to challenge the dominance of walled-off corporate systems with open-weight alternatives.

techcrunch.com/2025/07/15/mistral-releases-voxtral-its-first-open-source-ai-audio-model

Claude Code revenue jumps 5.5x as Anthropic launches analytics dashboard

Anthropic has launched a powerful analytics dashboard for its Claude Code AI assistant, giving engineering leaders real-time insights into developer productivity, tool usage, and ROI on AI coding investments.

venturebeat.com/ai/anthropic-adds-usage-tracking-to-claude-code-as-enterprise-ai-spending-surges

ChatGPT users send 2.5 billion prompts a day

ChatGPT receives 2.5 billion prompts from global users every day.

finance.yahoo.com/news/chatgpt-users-send-2-5-194443657.html

A ChatGPT ‘router’ that automatically selects the right OpenAI model for your job appears imminent

Like going to the supermarket and staring at aisles of cereal and sauces, the average ChatGPT user is currently faced with an overabundance.

venturebeat.com/ai/a-chatgpt-router-that-automatically-selects-the-right-openai-model-for-your-job-appears-imminent

Gabe Newell thinks AI tools will result in a 'funny situation' where people who don't know how to program become 'more effective developers of value' than those who've been at it for a decade

"Even if you're just a pure tool user you're going to find that the gains to utilising those tools are very, very high."

www.pcgamer.com/software/ai/gabe-newell-reckons-ai-tools-will-result-in-a-funny-situation-where-people-who-cant-program-become-more-effective-developers-of-value-than-those-whove-been-at-it-for-a-decade

Vibe coding service Replit deleted production database

: AI ignored instruction to freeze code, forgot it could roll back errors, and generally made a terrible hash of things

www.theregister.com/2025/07/21/replit_saastr_vibe_coding_incident

OpenAI and Anthropic researchers decry 'reckless' safety culture at Elon Musk's xAI | TechCrunch

The criticisms follow weeks of scandals at xAI that have overshadowed the company's technological advances.

techcrunch.com/2025/07/16/openai-and-anthropic-researchers-decry-reckless-safety-culture-at-elon-musks-xai

Small models, big wins: four reasons enterprises are choosing SLMs over LLMs

Why businesses are choosing SLMs over LLMs

www.techradar.com/pro/small-models-big-wins-four-reasons-enterprises-are-choosing-slms-over-llms

Using Gen AI for Early-Stage Market Research

Generative AI, particularly large language models (LLMs), offers a promising new tool for early-stage market research by simulating customer responses to product concepts.

hbr.org/2025/07/using-gen-ai-for-early-stage-market-research

OpenAI claims a breakthrough in LLM reasoning on complex math problems

OpenAI says its experimental language model has solved International Mathematical Olympiad (IMO) problems at a gold medal level—a possible breakthrough for AI with general reasoning skills. The results have not yet been independently confirmed.

the-decoder.com/openai-claims-a-breakthrough-in-llm-reasoning-on-complex-math-problems

Tool of the week: ChatGPT image styles

ChatGPT’s image tool got an upgrade that lets users select styles for the images. First select “Create image” under the tools menu.

Then you will see an option to select from nine pre-set image styles. ChatGPT will add a prompt that creates in the style you selected. You can then add to the prompt to create your image. Try this on the web version of ChatGPT (we didn’t see it in the app yet).

AI-generated image of the week

The WNBA All-Star game has basketball on our minds. And what is basketball without a signature shoe? Nike, call us!

Prompt: Design a signature basketball show that represents the brand Innovation Profs. Include the words "Innovation Profs" and have elements that represent being a college professor, generative artificial intelligence and innovation.

What we found

Common Sense Media's report, "Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions," found that 3 in 4 teens have used AI companions and more than half are regular users. Nearly one-third of teens (31%) find conversations with AI companions as satisfying or more satisfying than those with real-life friends.