In partnership with

NEW LAUNCHES

The latest features, products & partnerships in AI

• There’s a new open leaderboard just for Japanese LLMs

• Amazon robot pioneer aims to bring AI helpers to everyone

• Facebook Messenger adds AI backgrounds and Siri control to chats

• Microsoft partners with industry leaders on small specialized AI models

• AI music generator Suno v4 exceeds expectations

• Adobe’s new SlimLM model brings the power of AI to mobile devices

• Genies debuts AI-powered virtual worlds for brands and creators

• ChatGPT Advanced Voice Mode is here–Heres’ how to get it

IMPLEMENTATION

Announcements, strategies & case studies

• AI reshapes UI: 3 key steps for business leaders

• 7 methods to deploy a custom large language model

• Temenos partners with Nvidia to deliver on-premises AI for banks

• Thoughtworks’ latest Radar report says these are the AI solutions you should adopt right now

• AI in ESM: How smart organizations use AI to enhance operational efficiency

GOVERNMENT

Law, regulation, defense & politics.

• How Meta’s Llama AI model is being deployed to support sustainable development goals

• Trump pledges to reverse Biden’s AI policies amid global safety talks

• AI scholar Gary Marcus calls for new regulatory agency to oversee AI

• State Department launches task force to advance AI content authentication

• China hawks are causing an AI arms race, experts warn

• AI sovereignty debate intensifies as global adoption grows

• New US government taskforce will coordinate AI efforts on national security

IN OTHER NEWS

Compelling stories beyond the usual categories

• Microsoft is the mystery company licensing HarperCollins books for AI training

• Google AI uncovers 20-year-old software bug

• Nuclear power and AI: Is fission or fusion better?

• Why some industry insiders believe we’re a far cry from AGI

• Women to outnumber men in Gen AI use by 2025, research predicts

• Meta launches AI business unit under former Salesforce exec

• AI chatbots are turning into search engines and not everyone is pleased

What’s happening in AI right now

How infrastructure battles are reshaping the tech landscape

A battle for power is brewing - not just metaphorically, but quite literally. Industry forecasts suggest 40% of AI data centers may face power shortages by 2025, with total consumption projected to reach a staggering 500 Terawatt-hours by 2027. This isn't just another growing pain of the AI industry - it's sparking profound shifts in infrastructure, geopolitics, and corporate strategy. And beneath these infrastructure challenges lies an even deeper contest: the fight for AI sovereignty. Nations are increasingly viewing control over AI infrastructure not just as a technical necessity, but as a cornerstone of national security and technological independence.

The great power hunt

Major tech players are pursuing multiple paths to secure the energy they need. Equinix's new $260 million Singapore data center showcases one approach: building from scratch with sustainability in mind, featuring advanced cooling technologies and 100% renewable energy commitments. Meanwhile, others are getting creative with existing infrastructure - Patmos is converting a former Kansas City Star printing press into a $1 billion AI data center, leveraging existing power infrastructure for cost-effective development.

Nuclear renaissance

The search for reliable power is breathing new life into nuclear energy. Multiple nations have committed to tripling nuclear capacity by 2050, with tech companies increasingly exploring nuclear options for their data centers. While fusion technology generates excitement, fission remains the more practical near-term solution for meeting clean energy goals.

Infrastructure innovation

Companies aren't just seeking more power - they're revolutionizing how they use it. Cloudian's integration of Nvidia GPUDirect into object storage achieves 200GB/second throughput while reducing CPU usage by 42%. Schneider Electric has doubled production capacity at its Barcelona facility to meet surging demand for prefabricated data centers, offering rapid deployment solutions for companies racing to expand their AI capabilities.

Broader market forces

The infrastructure rush is reshaping the cloud market. Global cloud infrastructure spending reached $82 billion in Q3 2023, up 21% year-over-year, with AWS, Google, and Microsoft controlling 64% of total spending. This concentration of power raises questions about market access and competition, particularly as smaller companies and startups face significant barriers due to high computing costs and limited access to GPUs.

Technical headwinds

These infrastructure challenges come at a critical moment, as major AI companies struggle to achieve significant improvements in their next-generation models. OpenAI, Google, and Anthropic are all facing setbacks, suggesting that throwing more computing power at the problem may not be enough. The industry may need fundamentally new approaches beyond simply scaling existing technologies.

Political dimensions

The infrastructure landscape could shift dramatically based on political winds. A potential second Trump presidency could accelerate data center development through deregulation, while also affecting domestic semiconductor production and energy policies. Meanwhile, bipartisan support for the CREATE AI Act shows growing recognition of the need to democratize access to AI resources.

These developments point to a future where success in AI may depend as much on solving infrastructure challenges as on algorithmic breakthroughs. The winners will likely be those who can secure sustainable power sources, optimize their infrastructure for efficiency, and navigate an increasingly complex political landscape. The race isn't just about building better AI - it's about building the foundation to support it.

We publish daily research, playbooks, and deep industry data breakdowns. Learn More Here

One of the holy grails of unlocking the potential of AI is to build systems that can reason like humans. By improving AI's, Large Language Models in particular, ability to break down complex problems and apply logical steps.

Bagel’s research team has been exploring this problem. Analyzing LLM building techniques, especially fine-tuning techniques, to allow Large Language Models to evolve from pattern-recognizing prediction agents to true cognitive agents

Read More Here 🥯 

AI generated art

A look at the art and projects being created with AI.