• Today On AI
  • Posts
  • Sora hit 1M downloads faster than ChatGPT

Sora hit 1M downloads faster than ChatGPT

AND: Datacurve Raises $15M to Crowdsource Expert AI Training Data

TodayOnAI’s Daily Drop

  • Sora hit 1M downloads faster than ChatGPT

  • Reflection AI raises $2B to be America’s open frontier AI lab, challenging DeepSeek

  • Datacurve Raises $15M to Crowdsource Expert AI Training Data

  • 💬 Let’s Fix This Prompt

  • 🧰 Today’s AI Toolbox Pick

📌 The TodayOnAI Brief

Sora AI

🚀 TodayOnAI Insight: OpenAI’s video-generating app Sora has outpaced ChatGPT’s iOS debut, hitting 1 million downloads in under five days—despite being invite-only and iOS-exclusive. This marks one of the strongest app launches in the AI space to date.

🔍 Key Takeaways:

  • Sora saw 627,000 iOS downloads in its first seven days, surpassing ChatGPT’s 606,000 downloads during the same timeframe.

  • OpenAI’s Bill Peebles confirmed Sora hit 1 million installs in under five days, even with restrictions like invite-only access and North America–only availability.

  • Download momentum was immediate: Sora climbed to No. 1 on the U.S. App Store within four days of launch.

  • Daily installs peaked at 107,800 on October 1, maintaining strong performance throughout the week with lows still above 84,000.

  • Despite limited access, Sora is outpacing rival AI apps like Claude, Copilot, and Grok.

💡 Why This Stands Out: Sora’s explosive growth highlights surging consumer demand for generative video tools—even with friction like invite gates and limited platforms. The virality of Sora’s realistic deepfake outputs raises not just engagement, but ethical flags, as seen in growing backlash over AI-generated content of deceased celebrities. Will OpenAI’s content safeguards scale as fast as its adoption curve?

Reflection AI

🚀 TodayOnAI Insight: Reflection AI, a startup founded by ex-DeepMind researchers, has raised $2B at an $8B valuation—just seven months after being valued at $545M. The company aims to become the leading open source alternative to frontier AI labs like OpenAI, while countering China’s rapid advances in large-scale model training.

🔍 Key Takeaways:

  • Reflection AI is developing open-access frontier models trained on “tens of trillions of tokens,” with the first set to launch in 2026.

  • Its platform supports training massive Mixture-of-Experts (MoE) models—an architecture once limited to elite labs like OpenAI and DeepMind.

  • The 60-person team includes talent from OpenAI and DeepMind, and has secured a major compute cluster to support future development.

  • The company will open-source model weights while keeping datasets and pipelines proprietary—mirroring the approach of Meta and Mistral.

  • Revenue will come from enterprises and governments seeking customizable, sovereign AI systems with full infrastructure control.

💡 Why This Stands Out: Reflection AI is surfacing as a strategic counterbalance to both closed Western labs and open Chinese challengers like DeepSeek. Its hybrid “open but commercial” model signals a broader shift in how frontier AI will be built, shared, and monetized. As competition intensifies, the battle for global AI leadership may hinge less on algorithms—and more on access, sovereignty, and trust.

AI NEWS

🚀 TodayOnAI Insight: Datacurve, a Y Combinator alum, has raised a $15M Series A to rethink how high-quality training data is sourced—targeting software engineering tasks with a consumer-grade platform that turns expert data labeling into a rewarding experience. As the AI arms race for quality data accelerates, funders are backing novel approaches beyond the Scale AI playbook.

🔍 Key Takeaways:

  • Datacurve raised $15M in Series A funding led by Chemistry’s Mark Goldberg, with support from insiders at DeepMind, OpenAI, Anthropic, and Vercel.

  • The platform uses a “bounty hunter” model, paying expert software engineers for hard-to-source training data—over $1M in bounties distributed so far.

  • Co-founder Serena Ge emphasizes UX over payout: the platform is designed more like a consumer product than a traditional data labeling service.

  • The startup targets the rising need for complex post-training data, especially in RL environments used by cutting-edge AI models.

  • While currently focused on software engineering, Datacurve sees potential applications in other high-context domains like finance and medicine.

💡 Why This Stands Out: As foundational model performance hinges increasingly on tailored, domain-specific data, the market is shifting from scraping the open web to curating expert input. Datacurve’s hybrid approach—leveraging both financial incentives and UX design—signals a broader rethink of how AI-ready data is sourced at scale. Could the next great AI differentiator be not just the model—but who trains it?

💬 Let’s Fix This Prompt

 See how a simple prompt upgrade can unlock better AI output.

🔹 The Original Prompt

"Generate blog ideas for a tech company."

At first glance, this prompt might seem okay. But it's too broad — and that limits the quality of AI-generated results. Let’s improve it using prompt engineering best practices.

The Improved Prompt

Generate a list of unique, engaging blog post ideas for a B2B tech company that wants to attract decision-makers in mid-sized companies. Focus on topics related to emerging technology trends, industry insights, and practical solutions their software offers. Include suggested titles and a 1–2 sentence summary for each idea.

💡 Why It's Better

  • Specific audience: Targets decision-makers in mid-sized companies.

  • Contextual focus: Emphasizes emerging tech and practical solutions.

  • Actionable output: Requests summaries and titles to spark execution.

  • Tone and style: Guides the type of content (insightful, engaging, relevant).

🛠️ Learn how to adapt this prompt for SaaS, AI tools, dev teams & more →
Read the full PromptPilot breakdown

💡 Bonus Tool: Want to generate and master prompts instantly?
👉 Try PromptPilot by TodayOnAI (Free to use)

🧠 Smart Picks

📰 More from the AI World

  • Startup Battlefield company SpotitEarly trained dogs and AI to sniff out common cancers

  • Google ramps up its ‘AI in the workplace’ ambitions with Gemini Enterprise

  • OpenAI’s affordable ChatGPT Go plan expands to 16 new countries in Asia.

  • Zendesk says its new AI agent can solve 80% of support issues

🧰 Today’s AI Toolbox Pick

  • 🍔Sesame Labs (Community Tool): Creates community activities based on your website, social media, and Discord messages.

  • 💻ZipWP (Web Design Tool): Creates stunning websites in seconds.

  • 🍪Mr. Cook (Cooking Tool): A personal recipe generator tailored for your pantry.