AI Crafts Prompts, Meta Revamps Video
March 8, 2024: AI Self-Generates Prompts, AMD's China Setback, Meta's Video AI, AI Funding Boom, Adobe Goes AI Mobile, Reddit's Image Debate, Temasek Eyes OpenAI
Today's AIML news includes self-generated AI prompts, Meta's video revamp, AMD's regulatory hurdle, a surge in AI investments, Adobe's mobile innovation, Reddit's AI intrigue, and more.
1-Minute Roundup
AI surpasses humans in crafting prompts, enhancing creativity and efficiency.
US export controls thwarted AMD's AI chip project for China.
Meta aims to revolutionise video recommendations with advanced AI models.
Ideogram lands $80M for AI-driven image generation with clear text.
RapidSOS's Series C funding round amasses $150M, boosting public safety tech.
Jabali secures $5M seed funding for its AI-based game engine development.
Limbic raises $14M to expand its AI mental health services in the US.
Theia Insights gathers $6.5M to innovate the investment industry with AI analytics.
Healthee's Series A funding round brings in $32M for its healthcare platform.
Fluent obtains $7.5M for integrating LLMs with business databases and enhancing BI tools.
Zama's Series A funding round reaches $73M for homomorphic encryption technology.
Adobe introduces AI features to make content creation easier in its mobile app.
Reddit users debate over the authenticity of an AI-generated image.
Temasek considers a strategic investment in OpenAI.
Microsoft's AI image generator was criticised for creating explicit content.
Snowflake partners with Mistral AI to enhance its data cloud with AI.
Researchers develop ArtPrompt to challenge AI chatbot safety measures.
Carnegie Mellon and Los Alamos introduced AI to simulate chemical reactions.
IBM's SimPlan aims to improve AI planning capabilities by integrating language and traditional planning methods.
Spotlight
AI Prompt Crafting Shifts from Human Hands to AI Itself
Research reveals that AI models are better at creating their prompts, making the process faster and results more creative. This shift might change the role of prompt engineers, but human oversight remains crucial in integrating AI into practical applications.
AI's Creative Edge: Studies show AI-generated prompts outperform human-made ones, improving task and image outcomes.
Changing Roles: Automation in prompt generation suggests evolving prompt engineering roles, possibly towards broader AI management positions like LLMOps.
Evolving Interaction: As AI models advance, they may require less complex prompts, making interactions more straightforward and intuitive.
How do you see AI's increasing autonomy in tasks like prompt generation impacting human roles in AI development and application?
AMD's AI Ambitions Thwarted: US Curbs on China-Bound Chip
AMD's venture to tailor an AI chip for China hits a snag with US regulations, spotlighting the intricate dance of tech innovation and national security. As Washington tightens its tech export reins, the semiconductor industry braces for the ripple effects on global supply chains and market competition.
Regulatory Roadblocks: AMD's attempt to navigate US export controls faces challenges, requiring a special license for a chip designed to align with current restrictions yet deemed too advanced for China.
Market Movements: This development sends ripples through the stock market, with AMD shares dipping amid concerns over its competitive positioning against AI leader Nvidia.
Geopolitical Chess: The US continues its strategic clampdown on cutting-edge AI chip exports to China, aiming to curtail the flow of tech that could bolster military capabilities. This is reshaping tech exports.
How do you see the future of international tech collaborations and market dynamics evolving?
Meta's AI Makeover: Revolutionizing Video Recommendations
Meta is doubling on AI, unveiling plans to harness advanced models to revamp video recommendations across Facebook and Instagram. This ambitious leap, led by Facebook's head, Tom Alison, could redefine user engagement, making every scroll through your feed a tailor-made journey into digital content.
AI Integration: Meta's strategic pivot involves a unified AI model that will overhaul how videos are recommended, promising a more personalised and responsive user experience.
Efficiency Leap: The transition from product-specific AI models to a singular, robust architecture has already shown promise, with reel watch time witnessing a significant uptick.
Future Forward: Beyond videos, Meta's AI exploration extends to digital assistants and enhanced interaction features, signalling a broader commitment to integrating AI into the social fabric of its platforms.
What changes do you anticipate in how we discover and interact with content online?
AI Deals
Ideogram, founded by ex-Google staff specialising in AI image generation with legible text, secures $80M in funding.
RapidSOS closes a $150M Series C round, with $75M from BlackRock, bolstering its AI integration in public safety tech, raising its total funding to over $355M.
Jabali, an AI firm developing a generative AI game engine in San Francisco and Raleigh, secured $5M in Seed funding led by BITKRAFT Ventures with support from tech giants' angels.
Limbic, a mental health AI company based in NYC and London, secures $14M in funding led by Khosla Ventures to broaden its US healthcare reach.
Theia Insights, founded by an ex-Amazon Alexa scientist, raises $6.5M to revolutionise the investment industry with AI, led by Unusual Ventures.
Healthee raised $32M in Series A to enhance its employee healthcare platform with personalised, immediate support and guidance.
Fluent secures $7.5M to integrate LLMs with business databases, simplifying data queries and insights, aiming to transform traditional BI tool usage.
Zama, a Paris-based startup, raises $73M in Series A for its homomorphic encryption technology, reaching a valuation of nearly $400M.
Others
Adobe Revamps Mobile App with AI for Easier Social Content Creation
Adobe updates its Express app with Firefly AI, introducing features like "Text to Image" and "Generative Fill" for swift social media content design on mobile devices.
The app enhances workflow for Creative Cloud users and streamlines social media publishing. It is available in beta for Android and sign-up for iOS users.
Reddit's Latest Viral Debate: Is This Photo AI-Created?
Reddit users are engrossed in a debate over two images of a woman in a field. They are questioning which, if any, are AI-generated. This debate echoes past viral debates like the dress's colour. By the way, one image is actual, and one is AI-generated. Which is which?
The discussion highlights the growing difficulty in distinguishing between actual and AI-generated content and raises concerns about AI's impact on the future of creative jobs.
Temasek in Talks for Strategic Stake in OpenAI
Singapore's investment giant Temasek is negotiating a potentially groundbreaking investment in OpenAI, signalling state-backed confidence in AI's future.
Amidst OpenAI's rapid growth and ambitious chip-making ventures, these talks could fuel significant advancements despite regulatory scrutiny and high operational costs.
Microsoft Engineer Raises Alarm on AI's Dark Imagery
Microsoft's AI image generator, Copilot Designer, is under scrutiny for creating explicit sexual and violent visuals, and engineer Shane Jones is urging for corrective measures.
Despite internal reports and escalation, including letters to the FTC and Microsoft's board, the company faces challenges in mitigating content and copyright issues inherent in AI-generated imagery.
Snowflake Teams Up with Mistral AI to Boost Data Cloud AI.
Snowflake's partnership with Mistral AI brings top-tier language models to its platform, enabling advanced AI applications for users while ensuring data security.
This collaboration is backed by an investment from Snowflake Ventures, aiming to make AI technology more accessible and scalable for developers.
ArtPrompt, a New Tactic to Bypass AI Chatbot Safeguards
ArtPrompt, created by Washington and Chicago researchers, exploits ASCII art to sidestep safety measures in AI chatbots, prompting them to answer otherwise restricted queries.
The technique effectively disguises sensitive words as ASCII art, challenges current AI safety protocols and highlights the ongoing struggle to secure AI systems against misuse.
Carnegie Mellon and Los Alamos Develop AI for Simulating Chemical Reactions
Researchers have developed an artificial intelligence tool to predict how chemicals react, making experiments faster, cheaper, and more valuable for many science projects.
They're working to make this AI smarter to help discover new medicines and other science areas.
IBM's New SimPlan Makes AI Smarter at Planning
IBM researchers created SimPlan, a tool that helps AI better plan actions by combining language understanding with traditional planning methods.
SimPlan shows promise in solving complex tasks, improving AI's decision-making and problem-solving in various fields by blending the best of both worlds.
I wish you a wonderful weekend and look forward to our Monday catch-up!
