⚔️ Gemini Reigns Supreme 💔 OpenAI's Turmoil 🎨 FREE Midjourney Competitor
Your weekly dose of AI insights
Welcome to this week in AI.
This week, Gemini reigns supreme with its new experimental model, Midjourney has an open-source competitor and it’s better (free) and lastly, OpenAI has had a tumultuous time.
Let’s get into it!
👋 If you’re new here, welcome!
Subscribe to get your AI insights every Thursday (usually).
Black Forest Labs Launches FLUX.1, Open-Source AI Image Generator That Rivals Midjourney
Black Forest Labs just unveiled FLUX.1, a suite of cutting-edge AI models for image generation that are on par with industry giants like Midjourney and DALL-E 3.
In a remarkable feat, FLUX.1 surpassed the usage of Stability AI's SD3 model (released 3 months ago) in just three days, demonstrating its rapid adoption and popularity.
FLUX.1 comes in three versions:
FLUX.1 [pro]: The top-tier model offering unparalleled performance, accessible via API and for free on Replicate.
FLUX.1 [dev]: An open-source model matching the quality of the Pro version while being more efficient than competitors.
FLUX.1 [schnell]: An ultra-efficient model designed for personal use and local development.
Why This Matters
FLUX.1's impressive quality and open-source availability are set to shake up the AI image generation landscape, making high-quality tools accessible to a wider audience.
With Black Forest Labs also teasing their upcoming text-to-video model, which looks to rival industry leaders like Sora, we could potentially see a new wave of open-source, high-quality video generation tools emerging in the near future.
📰 Launch post by Black Forest Labs
📝 Analysis of Flux compared to Midjourney by Decrypt
Google's Gemini 1.5 Pro Surpasses Competitors, Leading the AI Race
Google DeepMind has made a notable leap in AI with its latest model, Gemini 1.5 Pro (Experimental 0801).
This large language model has outperformed top competitors like OpenAI's GPT-4 and Anthropic's Claude-3.5 Sonnet in various benchmarks.
Gemini 1.5 Pro excels in multilingual tasks, ranking first in languages such as Chinese, Japanese, German, and Russian.
It also demonstrates strong performance in technical domains, ranking 1st to 3rd in math, 1st to 2nd in instruction-following, 3rd to 5th in coding, and 2nd to 5th in hard prompts (English).
The model's competitive pricing and unique context caching feature, which can reduce costs by up to 75%, makes it an attractive option for businesses and developers.
Why This Matters
The release of the experimental Gemini 1.5 Pro signals that Google is finally catching up to the original perception of it’s abilities, Google fumbled consumer AI early on and is now a very serious player.
The new model has exceptional performance, particularly their multilingual capabilities and strength in technical domains, this in conjunction with competitive pricing positions it as a versatile tool for a wide array of enterprise applications.
⚔️ Play around with it in the Chatbot Arena
OpenAI's Tumultuous Times
OpenAI, a leading player in AI, finds itself at a crossroads, facing a series of challenges that could reshape its trajectory.
Leadership Exodus
The departure of co-founder John Schulman for rival Anthropic, following the previous exit of co-founder Ilya Sutskever, has significantly impacted OpenAI's leadership.
These departures, along with President Greg Brockman's extended leave, raise concerns about the company's ability to maintain its research and development momentum.
📰 Article by Tech Crunch
Elon Sues Again
Elon Musk has reignited his legal battle with OpenAI, alleging his former partners manipulated him into co-founding the company and then betrayed him by shifting from a non-profit to a for-profit model.
Musk's lawsuit, which he describes as a "textbook tale of altruism versus greed," claims "the perfidy and deceit is of Shakespearean proportions." This ongoing dispute could potentially tarnish OpenAI’s reputation.
📰 Article by The Guardian
Shifting Alliances
Microsoft's relationship with OpenAI has become increasingly complex.
Despite a long-term partnership and a reported $13 billion investment in the AI startup, Microsoft recently added OpenAI to its list of competitors in its latest annual report.
This move comes after OpenAI announced its own search engine prototype, SearchGPT, seemingly encroaching on Microsoft's territory.
Microsoft CEO Satya Nadella has expressed continued support for OpenAI, stating, "One of the things I love about Sam is every day he’s calling me and saying, ‘I need more, I need more, I need more’”.
OpenAI and Microsoft are increasingly competing for the same markets. At the same time, Microsoft is reducing its reliance on OpenAI's models by building up its own AI capacity, as demonstrated by the hiring of Mustafa Suleyman and the formation of Microsoft AI.
This strategic shift suggests a potential divergence in the relationship between the two companies, with Microsoft positioning itself as a more direct competitor to OpenAI.
📰 Article by CNBC
50% discount & Structured JSON Outputs
OpenAI has introduced Structured Outputs, a new feature in its API designed to enhance the reliability and consistency of AI-generated outputs.
This feature guarantees that model outputs strictly adhere to developer-supplied JSON schemas, ensuring that data is always formatted correctly.
Additionally, OpenAI has released the GPT-4o-2024-08-06 model, which offers substantial cost reductions compared to its predecessors.
Input costs have been slashed by 50%, now standing at $2.50 per million tokens, while output costs have been reduced by 33%, now at $10.00 per million tokens.
This model also boasts an increased context window of 16,384 output tokens.
📝 OpenAI's blog post
Character.AI and Google Strike Licensing Deal, Key Talent Returns to Google
Character.AI, the AI chatbot startup, has entered into a licensing agreement with Google, granting the tech giant access to Character.AI's LLM technology.
This move also sees Character.AI co-founders Noam Shazeer and Daniel De Freitas return to Google, where they previously led the development of the LaMDA language model.
As part of the deal, Character.AI will receive additional funding from Google.
Why This Matters
This strategic move by Google highlights a trend in the AI industry, where tech giants are acquiring licenses and talent from startups instead of outright acquisitions, avoiding regulatory scrutiny around mergers and acquisitions.
This is the latest in a string of similar deals made by Microsoft and Amazon, indicating a fierce competition for AI talent and technology.
Character.AI, which previously raised $193 million in venture capital and was reportedly getting 200 million monthly site visits, is expected to benefit significantly from this deal, with increased resources to further develop its AI chatbot technology.
It's also worth noting that Character.AI has open sourced Prompt Poet, an innovative tool designed to streamline and enhance prompt engineering in AI models.
Figure 02: The Next Generation Humanoid Robot
Figure 02, the latest humanoid robot from Figure, a California-based AI robotics company, boasts significant enhancements over its predecessor, Figure 01.
With upgraded AI, computer vision, batteries, and other components, Figure 02 is designed to perform a variety of tasks across commercial applications.
Notably, the robot utilises a Vision Language Model (VLM) for enhanced visual reasoning, enabling it to better understand and interact with its environment.
Why This Matters
Figure 02 represents a step forward in humanoid robot technology.
Its potential applications are vast, ranging from assisting in manufacturing environments to potentially even performing tasks in homes.
A partnership with BMW is already underway to test Figure 02's capabilities in automobile manufacturing, and the company anticipates the robot's deployment in customer sites by 2026.
This development places Figure 02 in direct competition with other humanoid robots like Tesla's Optimus, highlighting the growing race to bring advanced robotic assistants into our lives.
Quick bytes
OpenAI will give access to unreleased models to the U.S. AI Safety Institute
White House says no need to restrict ‘open-source’ AI for now
Fully-automatic robot dentist performs world's first human procedure
OpenAI built tool to detect AI generated text but won’t release it
Miss the last one?
That’s a Wrap!
If you want to chat about what I wrote, you can reach me through LinkedIn.
If you liked it, give it a share!