OpenAI reconfigure’s its Board of Directors
Plus Inflection AI releases a new model
Today’s Highlights:
📰 News: Inflection AI releases a new model + OpenAI reconfigure’s its Board of Directors
💰 Funding: Reddit seeks $5Bn+ IPO thanks to $200M+ in AI related revenue
⚡️ Top News Stories:
1.Inflection AI introduces Inflection-2.5, an upgraded in-house model that competes with leading LLMs like GPT-4 and Gemini, enhancing their personal AI, Pi.
Inflection-2.5 merges high intelligence capabilities with Pi's empathetic and helpful personality, aiming to provide a personal AI for everyone.
It achieved near GPT-4 performance levels using only 40% of the compute required for training, showcasing significant efficiency in model development.
Pi now includes world-class real-time web search capabilities, ensuring users receive up-to-date information and high-quality breaking news.
Since the rollout of Inflection-2.5, users have shown increased sentiment, engagement, and retention, contributing to organic growth.
2. OpenAI reinstates Sam Altman to its board, while adding 3 new members to enhance diversity and address governance criticisms.
Sam Altman has been reinstated to OpenAI's board of directors after an independent investigation by WilmerHale concluded his conduct did not mandate removal
The investigation, which involved interviews w/ board members and employees and a review of over 30K documents, found that Altman's firing was due to a breakdown in trust with the previous board, not concerns about product safety, security, or financial issues.
OpenAI announced the addition of three new board members: Sue Desmond-Hellmann, Nicole Seligman, and Fidji Simo, expanding the board to eight people and addressing previous criticisms regarding its all-male composition and lack of diversity.
OpenAI plans to adopt new corporate governance guidelines, including strengthening its conflict of interest policy and creating a whistleblower hotline, alongside establishing additional board committees focused on mission and strategy.
3. A U.S. govt.-commissioned report warns that AI development poses "extinction-level threats" to humanity, likening the risks to the introduction of nuclear weapons.
The report, produced by Gladstone AI, suggests radical policy actions, including making it illegal to train AI models with excessive computing power and creating a federal AI agency to regulate the industry.
It highlights the urgent need for "alignment" research to ensure advanced AI systems are safe and suggests tightening controls on AI chip manufacturing and exports.
The report reflects growing public concern over AI's potential dangers, with over 80% of Americans fearing AI could cause a catastrophic event.
Recommendations include potentially outlawing the publication of powerful AI models' "weights" and requiring government permission for training and deploying new models above certain thresholds.
4. Covariant, an OpenAI spinoff, has developed RFM-1, an AI model designed to enhance the capabilities of robots in performing tasks by learning from a diverse range of inputs including text, images, videos, robot instructions, and measurements.
5. Elon Musk's AI startup, XAI, plans to open-source its Grok chatbot, aligning with Musk's advocacy for more open and transparent AI development.
6. President Biden addressed the need for AI regulation in his State of the Union speech, specifically calling on Congress to ban AI voice impersonation.
7. Authors file class action against Nvidia, claiming NeMo AI was trained on unlicensed copyrighted works
8. Thomson Reuters plans to invest $8Bn in AI-focused acquisitions and its own AI development to transform its information services for professionals in legal, tax, and accounting sectors.
9. Midjourney has banned all Stability AI employees from using its service, alleging that their actions to scrape data caused a system outage.
10. Salesforce introduces Einstein Copilot: Health Actions, a conversational AI assistant designed to automate and streamline healthcare operations, enhancing clinical summaries and patient communication.
11. Apple experiments with AI to refine App Store ad placements, eyeing broader ad-supported service offerings.
💰 Top Funding News:
1. Reddit targets a $5.4Bn IPO valuation, banking on $200M+ in AI data licensing revenue to justify its ambitious share price range.
2. The AI2 Incubator, an initiative spun out of the Allen Institute for AI, secured $200M in compute resources for startups in its portfolio or program
The incubator focuses on accelerating the development of AI tech by offering startups up to $1M worth of dedicated AI-style compute, enabling them to train models beyond basic API options.
3. PayShepherd, which leverages AI within its contractor management solution to automate contract compliance and provide real-time cost reporting for industrial sites, raised a $7M CAD Seed round led by Inovia Capital.
4. Brevian, a no-code platform that enables business users in enterprises to build custom AI agents, particularly for support teams and security analysts, raised a $9M Seed Round led by Felicis.