Who's AI's Supreme Leader?

CONVERSATION STARTER

Nvidia Doubles Down on AI Supremacy with Blackwell, Rubin Chips!

Nvidia is leaving no doubt that it intends to maintain its edge in the AI chip wars. At Computex 2024, CEO Jensen Huang unveiled the company's roadmap for next-generation GPUs that will power increasingly advanced AI applications through 2026 and beyond.

The Blackwell Ultra is slated for release in 2025 and is what will likely springboard its current success even further. While full specs weren't provided, Nvidia says it will deliver a blistering 5x increase in teraflops performance compared to chips just two years prior. This unprecedented 2-year cadence of performance leaps is even outpacing Moore's Law.

Not resting on its laurels, Nvidia also teased "Rubin" for 2026 and "Rubin Ultra" for 2027 as part of what Huang dubbed "a new industrial revolution" powered by AI acceleration.

The audacious roadmap aims to maintain Nvidia's dominance in the AI training chip market, which has faced intensifying competition from AMD and others. AMD claims its forthcoming MI325X GPU, due in Q4 2024, will outperform Nvidia's current H200 flagship.

But Nvidia isn't just focusing on raw computing power. It also unveiled AI software tools like Project G-Assist for gaming assistance, and ACE for simplified creation of digital avatars across industries.

As AI capabilities rapidly advance, Nvidia is betting big that its full-stack dominance - from chips to software - will be pivotal in shaping the next era of artificial intelligence innovation.

PRESENTED BY TWILIO SEGMENT

What can good data do for you? Twilio Segment helps every team access good data. Data that's real-time, clean, and accurate. The result? Relevant customer experiences that drive real revenue.

Segment helps 25,000+ companies turn customer data into tailored experiences. With customer profiles that update in real-time, and best-in-class privacy features - Segment's Customer Data Platform allows you to make good data available to every team.

Good Data - Segment

MORE THAN MEETS THE EYE

Perplexity Launches AI-Powered 'Pages' for Automated Content Creation

Startup Perplexity has unveiled a new tool called Pages that allows users to automatically generate full articles, reports and guides using advanced language models.

With just a few prompts, Pages taps into Perplexity's AI search capabilities to find relevant information and synthesize it into a coherent, structured document complete with citations. Users can specify the target audience level and customize the writing style. But perhaps most impressively, Pages also automatically inserts relevant visuals and allows further editing of the content layout.

The burning question is whether AI-generatedPages will be adopted as legitimate content or dismissed as low-quality spam. Perplexity claims the tool is already getting indexed by Google, with the potential to disrupt search rankings and SEO strategies. It remains to be seen if machine-authored Pages provide true value or simply contribute to an oversaturation of AI word salads online.

Meta Using User Data to Train Its AI Models, Offers Opt-Out for Europeans

Privacy advocates, beware - Meta has confirmed it is using user data from Facebook and Instagram to train its AI models, including the AI systems behind features like chatbots and creative tools. And it's doing so without explicitly asking for consent, except in regions with strict data laws.

In Europe, where GDPR regulations provide robust data privacy protections, Meta must allow users to opt out of having their information utilized for AI training purposes. The company has begun emailing and notifying European users about this "AI at Meta" initiative, disclosing it will rely on a "legitimate interests" legal basis to train AI models on user content.

However, critics argue Meta's opt-out notices are deceptively worded, suggesting users need to justify their objection rather than it being an automatic right. The language states that a request will only be "honored" if the "objection is honored," implying Meta reserves the ability to deny some opt-outs arbitrarily.

Tech Giants Pledge Transparency on AI Safety Measures

In a move aimed at bolstering public trust, major tech companies including Microsoft, Amazon, and IBM have committed to disclosing the safety practices used in developing powerful AI models known as "foundation models."

At the AI Safety Summit in Seoul, 16 firms signed the "Frontier AI Safety Commitments," pledging to publish their processes for identifying and mitigating potential risks posed by large multimodal AI systems before deployment. The companies, including Meta, Samsung, and Anthropic, agreed not to release models if the risks cannot be properly controlled.

The commitments build on last year's "Bletchley Agreement" to classify AI risk categories. Signatories must outline their models' risk levels and safety roadmaps before the next AI Safety Summit in 2025. Government leaders touted the pact as establishing transparency best practices.

While welcomed by experts as a positive step, some argue the disclosures need third-party auditing to be fully credible. There are also concerns that AI companies may still exploit open-ended definitions of "acceptable risk" to their advantage.

HAVE YOUR SAY

ICYMI

• Elon Musk’s XAI raised $6 billion to build its supercomputer that will rival OpenAI and others.

• We can expect a new version of Siri which will allow voice control over individual apps.

• Arm Holdings shows off new chip designs that will be used for smartphones.

• Sam Altman may only have a $65,000 salary from OpenAI but has investment holdings worth at least $2.8 billion.

• A Swiss startup called Neuroplatform has created the world’s first bioprocessor which uses less power consumption than a digital chip.

MONEY, MONEY, MOOONEY

Funding rounds and acquisitions this week that caught our attention include:

• Zhipu AI raises $400 million from Saudi Aramco Venture Fund.

• AI translation platform DeepL raised $300 million last weekgiving it a $2bn valuation.

• Fintech player, Solutions by Text raised $110 million.

• Mavenir, a cloud-native network software builder, raised $75 million.

• Blockchain platform BabylonChain raised $70 million.

• Data privacy platform Transcend raised $40 million in a Series B.

LINKS WE’RE LOVIN’

• Tech gadget review: Sonos Ace Headphones by Marques Brownlee.

• Playlist of the week: Sunset Mix in Thailand by Chris Luno.

• Podcast of the week: The dangers of superintelligent AI with Roman Yampolskiy on the Lex Fridman show.

• Cheat sheet: Prompting checklicks for ChatGPT by Max Rascher.

• What to watch: We’re currently watching Tokyo Vice and loving it!

GIVE AI A GO

• AUTOBIOGRAPHER: It uses AI to tell your life story by taking your words and story and turning them into a written narrative.

• Eve Coach: AI-powered stress coach that can help replace stress with ease in the workplace or when you’re having a bad day.

• timeOS: AI-powered new tab page that provides contextual information from your apps to help you prepare your calendar.

• Monterey AI: Your AI copilot for product insights which can streamline feedback analysis and improve your customer success.

• Metaview: Automated, AI-generated note-taking with interview summaries designed for recruiters and interviewers.

STORIES BY AIJ CONTRIBUTORS

By creating content, demonstrating your expertise, and contributing, you can have your article in front of 100,000+ subscribers and get access to our 5 million+ yearly audience.Share your story todayfor free!

• 5 Ways AI is Driving Innovation Within The Energy Sector presented to you by Oil & Gas IQ.*

• 5 Ways of Enhancing Operations with IoT-integrated FSM Software by Mohd Sohel Ather.*

• Why scaling with AI is all about the human factor by Hannah Algar.

• Why React Native is the Perfect Choice for Cross-Platform Development by Erika Balla.

• GenAI: The new citizen developer in the room by Paul De Abreu.

• The Low-Code/No-Code Revolution: Democratizing Development And Reshaping Software Engineering by Denise Turley.

• AI could replace actors in just 5 years. But how likely is this? Written by Victoria Zavyalova.

• Building Enterprise Fluency in GenAI by Adam Lieberman.

• Keep your friends close and your employees closer: the rise of insider threats by Hannah Algar.

• How to have an open boardroom conversation on Generative AI by Solange Sobral.

• The Use Cases for AI are Starting to Snowball - and That’’s Good News for End Users by Jan Van Hoecke.

• Navigating the Ethics of AI Accessibility by Rosanne Kincaid-Smith.

• The key to balancing AI with human expertise by Andy Coussins.

• AI & Cybersecurity - how prepared should we be? Written by Martin Borrett.

• Unlocking the Potential: AI Optimisation in the Middle East by Martin Rosinski.

• Giving Your Chatbot Brand Personality in the Era of Generative AI by Frank Schneider.

• What do Lionel Messi and predictive modeling have in common? Written by Christina Kosmowski.

HOTTEST JOBS IN TECH

• Conversational AI Sales Specialist at Zendesk. This is a remote role.

• Senior AI Researcher at La Fosse. This is a hybrid role based out of London, UK.

• Technical Program Manager, Expert AI Trainer Acquisition and Engagement at OpenAI. This role is based out of San Francisco, USA.

• Director of AI at Alldus. This is a remote role.

• AI Consultant at Rabobank. This role is a hybrid based out of Utrecht, Netherlands.

TIME TO LEARN

• Deep Learning Specialization by DeepLearning.AI on Coursera. It is taught by course instructors Andrew Ng, Younes Bensouda Mourri, and Kian Katanforoosh

• AI For Everyone by DeepLearning.AI on Coursera. It is taught by course instructor Andrew Ng.

• Cloud Practioner by Amazon Web Services (AWS).