OpenAI’s Sora Leak, Perplexity's Hardware Move, Ai2's OLMo 2, and Uber's AI Push

Happy Wednesday, AI & Data Enthusiasts! From OpenAI’s leaked Sora video generator to Uber’s new venture into AI data labeling, there’s plenty to discuss. Plus, we have exciting updates on Ai2’s new model challenging Meta’s Llama and Perplexity's hardware expansion. Let’s explore!

In today’s edition: 

• OpenAI’s Sora Video Generator Leaks Amid Artist Backlash

• Perplexity Ventures Into Hardware

• Ai2 Challenges Meta’s Llama with OLMo 2

• Uber Enters the AI Annotation Arena

- Naseema Perveen

WHAT CAUGHT OUR ATTENTION MOST

OpenAI’s Sora Video Generator Leaks Amid Artist Backlash

In a dramatic turn, OpenAI’s video generator, Sora, appears to have been leaked by a group of testers. The group, calling itself “Sora PR Puppets,” claims the leak is a protest against OpenAI’s handling of its artist program and transparency issues.

• Leaked Access via Hugging Face: The group published a project on Hugging Face that used OpenAI’s Sora API to generate 10-second, 1080p videos with a simple text prompt. The front end reportedly relied on authentication tokens from OpenAI's early access program.

• Artists Push Back: The group alleges OpenAI pressured early testers to promote Sora positively while providing inadequate compensation for their labor. They accuse OpenAI of prioritizing PR over meaningful artist feedback.

• OpenAI Responds: OpenAI confirmed Sora remains in “research preview,” emphasizing that participation is voluntary and that artists are offered free access, grants, and other forms of support. The company aims to balance creativity with safety measures.

• Challenges Loom for Sora: Despite its potential, Sora faces stiff competition from rivals like Runway and Stability, both of which have secured high-profile partnerships. Technical issues, including consistency and speed, have also hindered Sora’s progress.

This leak raises questions about OpenAI’s approach to collaboration and transparency. As the company refines Sora and prepares for a broader launch, it will need to address both technical hurdles and trust issues with its creative community. Whether Sora can regain momentum and establish itself in the competitive video generation space remains to be seen.

IN PARTNERSHIP WITH PINATA

Streamline your development process with Pinata’s easy File API

• Easy file uploads and retrieval in minutes

• No complex setup or infrastructure needed

• Focus on building, not configurations

Try today!

WORTH WATCHING

In this eye-opening video, former Google X executive Mo Gawdat shares a dire warning about the rapid advancement of artificial intelligence. Drawing on his deep industry knowledge, Mo discusses the potential risks of unchecked AI development, from ethical dilemmas to existential threats. This urgent message challenges us to rethink our approach to AI governance and responsibility.

Watch now to understand why Mo believes the time to act is NOW!!!

KEEP YOUR EYE ON IT

Perplexity Ventures Into Hardware

Perplexity, the AI-powered search engine, is exploring the idea of entering the hardware space. Founder and CEO Aravind Srinivas recently teased a low-cost device designed for voice-to-voice Q&A interactions, sparking both excitement and questions about its feasibility.

• A Viral Idea Gains Momentum: On Monday, Srinivas posted on X about creating a "simple, under $50" device that could reliably answer user questions. He promised to move forward if the post received over 5,000 likes. After achieving that milestone, his response — “Alright. LFG!” — suggests Perplexity is serious about pursuing this innovative idea.

• Rising Interest in AI Hardware: Perplexity is not alone in this pursuit. MidJourney formed a hardware team earlier this year, and OpenAI’s Sam Altman is reportedly collaborating with Jony Ive on a next-gen AI device. The hardware push reflects the tech industry’s belief that physical AI tools can unlock new forms of interaction and engagement.

• A History of Challenges in AI Devices: Despite the buzz, the hardware journey is fraught with risks. For instance, Rabbit’s R1, a popular AI device, struggled to deliver promised features, leading to steep discounts on secondary markets. Humane’s Ai Pin, a wearable touted as a smartphone replacement, flopped due to weak sales and safety issues, forcing recalls and putting the company on shaky ground.

• The Role of Funding and Execution: Perplexity’s strong financial position — with rumors of a half-billion-dollar funding round on the horizon — gives it a crucial advantage. However, as history shows, AI hardware success requires more than money. The company must navigate technical challenges, deliver on user expectations, and ensure flawless execution to make this device a success.

This potential move into hardware could mark a turning point for AI accessibility, making intelligent tools more affordable and user-friendly. 

Ai2 Challenges Meta’s Llama with OLMo 2

The Allen Institute for AI (Ai2) has unveiled OLMo 2, a family of truly open-source language models designed to rival Meta’s Llama series. These models push the boundaries of openness, transparency, and performance in AI development.

• Open from Start to Finish: The OLMo 2 models meet the Open Source Initiative's AI definition, ensuring complete transparency. They include open-access training data, code, recipes, intermediate checkpoints, and evaluations.

• Performance and Scale: The OLMo 2 family consists of two models: OLMo 7B (with 7 billion parameters) and OLMo 13B (with 13 billion parameters). These models were trained on a 5 trillion-token dataset, which includes academic papers, Q&A forums, math workbooks, and curated websites. They demonstrate competitive performance, even outperforming Llama 3.1 8B in benchmarks.

• Commercial-Friendly License: OLMo 2 models are released under the Apache 2.0 license, allowing for both academic and commercial applications.

OLMo 2’s release comes amid debates over the safety of open models, as incidents of misuse, such as Llama's adoption of defense tools, highlight risks. Ai2 argues that openness fosters transparency, reduces concentration of power, and drives ethical advancements. With its focus on accessibility and innovation, OLMo 2 aims to empower developers worldwide while setting a new standard for open-source AI.

Uber Enters the AI Annotation Arena

Uber is taking its gig economy expertise into the world of AI by building a workforce dedicated to data labeling and annotation. With a new division and ambitious plans, the ride-hailing giant is gearing up to support both internal and external AI projects.

• Introducing Scaled Solutions: Uber’s newly launched AI annotation and data labeling division will support internal units and external clients, including Aurora Innovation and Niantic.

• Global Workforce Expansion: The company is recruiting contractors in multiple countries, including the U.S., Canada, and India.

• Riding the AI Wave: Data labeling is a booming market, driven by AI advancements. Competitors like Scale AI recently raised $1 billion at a $13.8 billion valuation, underscoring the sector's potential.

Uber’s entry into data labeling signals its ambition to play a bigger role in the AI ecosystem. By leveraging its global gig workforce model, the company is well-positioned to tap into this growing market.

TERM OF THE DAY

Data Wrangling

Data wrangling means cleaning and organizing messy data so it can be used easily. Imagine you have a list of names, but some are misspelled, some are in all caps, and others have extra spaces. Fixing these issues—like correcting spellings, making the text consistent, and removing spaces—is data wrangling. It helps turn confusing data into something clear and useful!

Read more terms like this in our Glossary. 

TOGETHER WITH THE RUNDOWN AI

Learn how to make AI work for you

AI won’t take your job, but a person using AI might. That’s why 800,000+ professionals read The Rundown AI – the free newsletter that keeps you updated on the latest AI news and teaches you how to use it in just 5 minutes a day.

Sign up to start learning.

ICYMI

• Nvidia CEO Jensen Huang claims trustworthy AI is still years away. 

• Anthropic unveils a new standard for AI and data integration

• OpenAI is funding AI morality research.

• Apple reportedly developing conversational Siri using LLMs.

• Ai2’s new open source Tülu 3 lets anyone play the AI post-training game.

$$$ MONEY MATTERS

• Juna.ai raised $7.5 million for its AI platform for industrial process control.

• New Lantern secured $19 million for its AI radiology resident.

• Blue Bear Capital raised $200 million for AI climate and energy bets.

• AI data center startup Crusoe is raising $818 million to expand its operations. 

• Google is committing $20 million to research using AI to advance science.

LINKS WE’RE LOVIN’

✅ Podcast: AI for Science with Sir Paul Nurse, Demis Hassabis, Jennifer Doudna, and John Jumper. 

✅Cheat sheet: Cheat Sheets for AI, Neural Networks, Machine Learning, Deep Learning & Data Science PDF.

✅Course: Deep Learning Specialization. 

✅Whitepaper: The Power Of Your Unstructured Data With NLP.

✅Watch: OnePlus 13 Full Review. 

A Quick Question before you go…

SHARE THE NEWSLETTER & GET REWARDS

Your referral count: 0

Or copy & paste your referral link to others: https://aijournal.beehiiv.com/subscribe?ref=PLACEHOLDER

That’s all for now. And, thanks for staying with us. If you have specific feedback, please let us know by leaving a comment or emailing us. We are here to serve you!

Join 130k+ AI and Data enthusiasts by subscribing to our LinkedIn page.

Become a sponsor of our next newsletter and connect with industry leaders and innovators.