- Business Bloopers
- Posts
- The Pentagon's AI drone initiative
The Pentagon's AI drone initiative
Business Bloopers
December 07, 2023
Greetings, fellow humans. 👋 Hope you're doing awesome! A lot of exciting things have happened in the past week. From autonomous military Weapons to ChatGPT’s new biggest competitor, I have got you covered on the best.
🧵 In today's edition:
⚔️ Autonomous weapons
🎬 What people are calling the “ChatGPT For Video”
👋 Amazon's AI chatbot, Q, leaking confidential information
❤️ My favourite AI tools this week
🌍 AI News Roundup
🐦 That’s a wrap
Read Time: 4 min 39 sec
LATEST ADVANCEMENTS
Image Source: Microsoft Bing
The Briefing: The Pentagon is ramping up its "Replicator" initiative, gearing up to unleash a fleet of lethal, autonomous AI-driven vehicles by 2026 in a race to keep pace with China's strides.
What's happening:
"Replicator" is all about rapidly expanding AI attack and defense drones.
Even though the U.S. military insists on human control, experts can't shake off concerns about the looming inevitability of full autonomy.
There's a growing worry that the breakneck speed of AI development might leave the military struggling to thoroughly test systems before putting them into action.
Why it matters: Merging lethality with autonomy raises red flags, calling for vigilant oversight. The swift evolution of AI in warfare brings pressing moral questions to the forefront, emphasizing the critical need for thoughtful consideration.
Read more: Yahoo News
✨Google DeepMind reveals Gemini
The Briefing: Google DeepMind has unveiled three versions of its most advanced multimodal AI yet — Google Gemini, (AKA, GPT-4’s biggest competitor)
Gemini is available in three sizes: Ultra for intricate tasks, Pro for versatile applications, and Nano for efficient on-device operations.
Gemini Ultra excels in 30 out of 32 benchmarks used in LLM research & development, achieving a groundbreaking 90% on the MMLU benchmark, surpassing human expert performance.
With cutting-edge features like sophisticated reasoning, multimodality, and advanced coding, Gemini represents a new era in AI capabilities.
Running on Google's latest Cloud TPU v5p, Gemini Pro is integrated into Google products like Gmail, YouTube, and Docs through Bard, available in English in over 170 countries.
Gemini Ultra is set to launch early next year, and Gemini Nano will be directly available on Pixel 8 devices.
Why it matters: Google's highly anticipated 'ChatGPT killer,' Gemini, is now a reality, offering three distinct models for enhanced flexibility. While the full potential of Gemini Ultra is yet to be tested, users can experience Gemini Pro in Bard, currently recognized as the top free chatbot on the market based on benchmarks.
The Briefing: In the cutting-edge realm of generative AI, Pika is turning heads and making waves, securing a remarkable $55 million in funding led by Lightspeed Venture Partners. Their revolutionary platform, "Pika 1.0," is not just an editing tool – it's a game-changer, allowing users to effortlessly create diverse styles, from 3D animation to anime, using simple captions and images.
Revolutionising the Game: Pika 1.0 is setting new standards in the AI landscape, extending video length and enabling users to change styles with just a few clicks. While industry giants like Google and Meta delve into generative AI, Pika stands out, capturing the attention of industry leaders and investors alike.
Empowering Creativity: Guided by former Stanford PhD students, Pika boasts a solid technical foundation. Beyond technology, Pika is about empowering users to bring their creative visions to life, turning them into professional-quality videos. In the ever-evolving AI landscape, Pika isn't merely keeping pace – it's at the forefront, leading a transformative journey in visual creativity.
The Briefing: Q was designed to be a more reliable alternative to consumer-focused AI chatbots but Amazon employees say the chatbot is leaking confidential data and even experiencing "severe hallucinations," according to Platformer.
Story by Lakshmi Varanasi.
The Details:
Leaked communications: Q may reveal confidential info, like AWS data centers and unreleased features.
"Severe hallucinations": Q confidently provides inaccurate info, especially in legal advice.
Amazon denies security issues, refining Q for general availability.
Challenges highlight complexities in developing secure AI assistants, akin to issues faced by Microsoft's Sydney
My Thoughts:
The alarming prospect of leaking confidential information raises significant concerns. The potential risks associated with inaccuracies, especially involving sensitive data, underscore the crucial need for stringent security measures to ensure user trust and data integrity.
❤️ My Favourite AI Tools this Week
💼 Coconaut- Create a chatbot like ChatGPT powered by your data (link)
💡 PromptForge- Prompt extension for ChatGPT, Bard & Claude (link)
📈 Ventureki- AI business plan generator (link)
🧠 Blaze- Web3 CRM, Marketing automation and growth (link)
🎓 Learniverse- AI-powered personal learning revolution (link)
✈️ Itair- AI-powered travel planning for human-created memories (link)
🖥️ Super-Create a website from your Notion database in minutes (link)
AI NEWS ROUNDUP
EnCharge AI has just secured an impressive $21.7 million in venture funding to enhance its AI accelerator hardware. Fuelled by in-memory computing, this technology not only elevates AI applications but also champions environmental friendliness by significantly reducing power consumption. The aim is clear: offer high-quality, budget-friendly, and eco-conscious AI solutions, addressing the cost and energy challenges posed by current AI chips.
The AI Governance Alliance aims to design transparent and inclusive AI systems, recognizing both the promise and perils of generative AI. The initiative seeks to ensure that AI is always in service to society, emphasizing the importance of collaboration and responsible AI development.
MIT, Harvard University and University of Washington researchers introduce HuGE, a breakthrough in AI learning that skips the need for expert-crafted rewards. Instead, it taps into feedback from everyday folks worldwide, making the learning process faster and more scalable. HuGE outshines other methods, proving its potential for personal and efficient AI training in real-world scenarios.
During a Russian conference last Friday, Vladimir Putin warned that the West can't monopolize AI. He pledged that Russia would intensify both funding and aspirations to narrow the AI gap between the country and the US and China.
THAT’S A WRAP
I hope you guys enjoyed this Newsletter. I am always wanting to improve so If you have a second, I would love to hear what you guys think of this newsletter. I have enabled comments on my website Here, or you could hit reply to this email. I would love to hear feedback from you guys to see what you want to see from me in the future and what you like.
I am currently experimenting with new styles of newsletters, so if you liked the look and format of this email more than the last one, I would love to hear what you think.
Keep Embracing the Future, One Byte at a Time! 🔄
See You Next Week
Reply