Good morning, it's Tuesday! This week, Apple preps Siri for an AI makeover, Cloudflare gets serious about bots paying their dues, and LinkedIn quietly uses your data to train its AI (thanks?). Meanwhile, ByteDance is cooking up its own AI chips, Meta’s smart glasses are somehow cool, and Three Mile Island’s back—this time, to power Microsoft’s data centers. Oh, and Jony Ive and Sam Altman are teaming up for a mystery AI gadget. Grab your coffee, grab you spectacles, let’s jump in.
Your Daily Roundup:
Three Mile Island Nuclear Plant to Reopen, Power Microsoft Data Centers by 2028
Cloudflare to Launch Marketplace for Websites to Charge AI Bots for Scraping
Jony Ive Confirms AI Hardware Collaboration with OpenAI's Sam Altman
Apple begins testing AI software aimed at making Siri smarter on the iPhone 16.
LinkedIn Uses User Data to Train AI Models, Requires Two Opt-Out Steps
ByteDance Taps TSMC to Develop AI GPUs, Aiming to Reduce Reliance on Nvidia
Meta's Ray-Ban Smart Glasses Succeed Where Other AI Gadgets Failed
👉️ Top AI Stories
Three Mile Island Nuclear Plant to Reopen, Power Microsoft Data Centers by 2028
Three Mile Island, the Pennsylvania nuclear plant infamous for a 1979 reactor meltdown, will be reopened by Constellation Energy to power Microsoft’s data centers, supporting its cloud and AI operations. Renamed the 'Crane Clean Energy Center', the plant will provide carbon-free energy under a 20-year agreement and is expected to create 3,400 jobs and contribute $16 billion to Pennsylvania's GDP. → Continue reading here.
Cloudflare to Launch Marketplace for Websites to Charge AI Bots for Scraping
Cloudflare announced plans to introduce a marketplace next year allowing website owners to charge AI model providers for scraping their content. In response to growing concerns about AI bots using website data without compensation, Cloudflare’s CEO emphasized the need for creators to be compensated to sustain content creation. Additionally, Cloudflare has introduced AI Audit, a tool giving websites control over AI bot access, letting owners block or permit specific scrapers. → Continue reading here.
Jony Ive Confirms AI Hardware Collaboration with OpenAI's Sam Altman
Jony Ive, the former Apple design chief, has confirmed he's working with OpenAI CEO Sam Altman on a new AI hardware device, as revealed in a profile in The New York Times. Though details are scarce, the project, currently involving just 10 employees, including key iPhone collaborators, is being led by Ive's company, LoveFrom. The venture is exploring how generative AI could create a groundbreaking computing device that surpasses traditional software capabilities, with potential for up to $1 billion in funding by year’s end. → Continue reading here.
Apple begins testing AI software aimed at making Siri smarter on the iPhone 16.
Apple has released a test version of its latest software update, incorporating "Apple Intelligence" AI features into Siri, which will automate tasks and enhance user experience on the new iPhone 16 models, launching Friday. The update, currently available for testing, will initially support U.S. English, with plans to expand to other regions and languages next year, and aims to differentiate Apple's AI offerings from competitors like Samsung and Google. → Continue reading here.
LinkedIn Uses User Data to Train AI Models, Requires Two Opt-Out Steps
LinkedIn has quietly opted users into allowing their personal data to be used for training generative AI models, as revealed by recent updates to its privacy policy. While the platform claims it uses privacy-enhancing technologies to remove personal data from training sets, users can opt out by adjusting their 'Data for Generative AI Improvement' settings and submitting a 'LinkedIn Data Processing Objection Form'. However, opting out only stops future data use, not past data already used for AI training. → Continue reading here.
ByteDance Taps TSMC to Develop AI GPUs, Aiming to Reduce Reliance on Nvidia
ByteDance, TikTok’s parent company, is developing two AI GPUs for AI training and inference, to be produced by TSMC by 2026. This move follows ByteDance’s heavy spending—over $2 billion on Nvidia GPUs this year alone—and aims to address the high costs and shortages of Nvidia's hardware. The shift could help ByteDance lower costs, though challenges include creating a compatible software stack, as the company currently relies on Nvidia’s CUDA for AI operations. → Continue reading here.
Meta's Ray-Ban Smart Glasses Succeed Where Other AI Gadgets Failed
Meta's Ray-Ban smart glasses have defied expectations in a year when most AI wearables struggled, succeeding due to their stylish design, affordability, and practical features like audio and video capture. Unlike other AI gadgets that overpromised and underdelivered, Meta's glasses incorporate AI as a helpful tool rather than the main focus, allowing users to identify objects, translate text, and make video calls. → Continue reading here.
☝️ Sponsor: Weights & Biases
Make real progress on your LLM development, click here to get started with Weave today.
👾 Forward Future Original
This article is the first in a 4-part Forward Future Original series by guest author Kim Isenberg.
Part 1: What is AGI, anyway?
Introduction
AGI is generally regarded as the holy grail in the AI scene. Along with ASI and the singularity, it is considered the most far-reaching and world-changing moment in that it is the necessary prerequisite for the others in this chain. It will be a decisive moment because AGI will have an influence and impact on all areas of human life. (Narrow) artificial intelligence (i.e. the currently common language models) has already gradually found its way into the diverse areas of people's lives and work and is widely used in a wide range of areas. For example, excellent translations no longer have to be painstakingly done by professional translators, but are automatically translated almost perfectly by AI. E-mails can be written and sent with a simple prompt. 41% of all the code flourishing on the internet is already AI-generated1 , and first-level support is now provided by AI instead of real people (the latest example being the fintech Klarna)2 .
AGI, in turn, will significantly build on this. The accuracy and generalizability achieved by applying the eponymous Artificial GENERAL Intelligence will, on the one hand, create even broader application as the generality encompasses even more fields of work, and, on the other hand, better output through more precise reasoning. In addition, it is unanimously accepted that AGI can also act and operate largely autonomously:
It is widely accepted that an AGI is a highly autonomous artificial system that has at least the same or better cognitive and intellectual abilities than humans in all areas. It understands and performs any intellectual task that a human can understand and perform.
🚀 Launches + Funding
Microsoft Expands AI with Copilot Updates: Microsoft's second Copilot update adds new AI tools like Copilot Pages, agents, and Python integration in Excel to boost productivity across 365 apps.
Black Forest Labs Raises $100M: Generative AI startup Black Forest Labs raises $100 million at a $1 billion valuation, advancing its innovative image generation technology and text-to-video development.
Letta Launches with $10M for AI Memory: Letta, backed by $10 million in seed funding, is developing MemGPT to address AI memory limitations, aiming to enhance long-term interactions with stateful AI agents.
✌️ Sponsor: VULTR
Vultr is empowering the next generation of generative AI startups with access to the latest NVIDIA GPUs. Try it here and use promo code "BERMAN300" for $300 off your first 30 days.
✍️ Editor Picks
Research
AI Model Identifies Whale Species and Sounds for Conservation
Google Research has launched a new whale bioacoustics model capable of identifying vocalizations from eight distinct whale species, including Bryde's whales' unique "Biotwang" sound. This model, developed to aid conservation efforts through long-term passive acoustic monitoring, analyzes underwater recordings, helping researchers track and study elusive whale populations in remote environments. → Continue reading here.
Models
Qwen 2.5 72B Tops Open Weights Leaderboard, Surpassing Larger Models in Speed and Efficiency
Qwen 2.5 72B, released by Alibaba, has claimed the top spot on Artificial Analysis' Quality Index, outperforming even the larger Llama 3.1 405B in coding and math tasks. Despite being smaller in size, Qwen 2.5 boasts faster performance and competitive capabilities, featuring a 128k context window and 8k output tokens, double that of Llama 3.1. Priced at $0.4 per million input and output tokens, it is approximately 10 times cheaper than GPT-4o, offering a cost-effective solution for AI applications, with Hyperbolic Labs and DeepInfra already adopting the model. → Continue reading here.
Health
AI Meets Fruit Fly, Creating Efficient Brain Models
Researchers have developed a virtual brain model based on the fruit fly’s visual system, predicting neuron behavior with striking accuracy and offering a powerful tool for brain science and AI. This biologically inspired approach suggests AI systems could become significantly more energy-efficient by mimicking the compact, efficient neural strategies found in the fly’s 100,000 neurons, which perform complex tasks while consuming far less power than today's AI technologies. → Continue reading here.
🧰 AI Toolbox
KLING AI 1.5 Boosts Speed and Creative Features: The KLING AI 1.5 update offers faster video generation, 1080p HD videos in Professional Mode, and new tools like the Motion Brush for custom video motion paths. It also supports simultaneous video creation, longer 10-second videos, and improved image resolution, with a new user guide for support.
HeyGen Offers AI Video Creation in 175 Languages: HeyGen enables users to create and translate studio-quality videos in 175 languages using AI, making video production accessible without a camera or team. The platform offers unlimited video generation and is available for free to start.
🛰️ Houston, We Have More Headlines!
AI's Future Transformations: AI will revolutionize fields like education, driving progress while addressing global challenges and societal impacts.
Meta's New Personalized Image Model: Meta's "Imagine Yourself" model generates personalized images without tuning.
OpenAI Launches Academy for AI Developers: OpenAI Academy supports developers by providing resources to accelerate innovation.
Google's Dataset Tracks Urban Growth Globally: Google tracks building changes and heights across the Global South from 2016 to 2023.
Amazon Unveils AI Assistant for Sellers: Amazon's AI assistant Amelia helps third-party sellers manage account issues and access data.
UN AI Governance Recommendations: A UN advisory body recommends a global AI panel, standards exchange, and transparency.
China's AI Firms Adapt to Chip Bans: Chinese AI companies innovating around chip sanctions by optimizing software to maintain progress.
Help Us Improve
Are you enjoying Forward Future’s newsletter? |
Reply to this email if you have specific feedback to share. We’d love to hear from you.
Stay Connected
Looking for more AI news, tips, and insights? Follow us on X for quick daily updates and bite-sized content.
For in-depth technical analysis, subscribe to the Forward Future YouTube channel. We dive deep into new models, test their performance, explore the latest tools, and share our impressions of AI innovations and developments.
Prefer using an RSS feed? Add Forward Future to your feed here: RSS Link
Thanks for reading this week’s newsletter. See you next time!
🧑🚀 Forward Future Team
