This Week’s BIG thing

ChatGPT’s rival for Google is finally here

The hotly anticipated ChatGPT search has finally been released by OpenAI. This is an upgrade that integrates real-time web search into the popular AI chatbot. Here’s what it offers:

Key Highlights

ChatGPT Search enables users to access up-to-the-minute information directly from the web. Key features include:

• Instant (very fast) updates on news, sport, stocks, and weather.

• A conversational interface that feels natural and intuitive.

• Source links for transparency.

• An ad-free browsing experience.

• Visuals such as interactive graphs for a richer search experience.

Access and Rollout

The rollout is being phased in:

• Available now for ChatGPT Plus and Team subscribers.

• Enterprise and education users will get access soon.

• Free-tier users will join in the coming months.

• Accessible on chatgpt.com and through mobile or desktop apps.

Technology Behind the Feature

The search capabilities are built on:

• A specialised version of GPT-4.

• Integration with search engines like Microsoft Bing.

• ·Collaborations with major publishers and news sources.

Planned Improvements

OpenAI plans to further enhance ChatGPT Search with:

• Improved shopping and travel search options.

• Integration of Advanced Voice Mode.

• Enhanced research tools using new AI models.

• Personalised search results based on user preferences.

Comparison with Google and Perplexity

ChatGPT Search aims to challenge Google’s dominance, leveraging its conversational and ad-free experience to attract users. Although Google still holds a massive market share, early feedback favours ChatGPT for its personalised and streamlined approach. We are big fans of Perplexity and I wouldn’t say that ChatGPT search offers anything significant over Perplexity other than its superfast outputs.

Early perception

GPT-based search tools like ChatGPT and SearchGPT are creating a mix of excitement and caution. Users are enthusiastic about the conversational, intuitive nature of GPT search, which offers synthesised rapid answers instead of just links, making the process feel more like a helpful dialogue. However, traditional search engines like Google still dominate due to their reliability and familiarity, leading many to hesitate in fully embracing GPT-based search for critical tasks.

Tech enthusiasts are exploring GPT tools, but general users are slower to switch, preferring the interfaces they are already comfortable with. Media and public opinion remain divided—some view GPT search as a revolutionary development, while others are concerned about misinformation and the opaque nature of AI-generated results.

The Rabbit hole 🐰

• _{Hear from OpenAI themselves} 🔗

• _{Watch a further comparison of ChatGPT search} 🔗

• _{Can ChatGPT search take on Google?} 🔗

• _{How Perplexity works} 🔗

What’s new: Tech

5 Key Takeaways from OpenAI's Ask me Anything Dialogue on the Future of AI

5 Key Takeaways from OpenAI's Ask me Anything Dialogue on the Future of AI

OpenAI has been making significant strides recently, particularly with the release of enhanced features in ChatGPT, such as search and multimodal capabilities. In a recent Reddit ask me anything (AMA), Sam Altman and key OpenAI leaders shared valuable insights into the current trajectory of AI development.. Here’s a quick breakdown of what we learned, along with some thought-provoking critiques:

1. AGI with Today’s Hardware

OpenAI is focused on making architectural and algorithmic advances to achieve Artificial General Intelligence (AGI) without waiting for revolutionary new hardware. They are pushing current tech to its limits through efficiency improvements and better model design, suggesting either more opportunity to be had with the current wave of tech, or that compute challenges are driving this approach.

2. Multimodal “O” Series Models Instead of GPT-5

OpenAI seems to be moving away from larger language models, opting instead for enhancing “o” series models that can handle multimodal inputs—like text and images—and feature autonomous capabilities. This marks a shift towards creating versatile AI that interacts with the world in richer, more meaningful ways. But some critics argue this could indicate a strategic pivot to mask the fact that simply scaling up models is yielding diminishing returns. Is OpenAI pivoting because they’ve realised bigger isn’t necessarily better, or are there underlying limitations preventing a GPT-5?

3. Compute Challenges

Compute is becoming a scarce resource, akin to a new form of currency in AI. OpenAI highlighted that future AI development will need to innovate on how compute is organised and distributed, potentially reshaping the entire infrastructure that supports AI development. However, this raises a critical question—how sustainable is this reliance on massive compute in the long term? Are we building an AI future that only a handful of companies can afford to sustain, thus centralising power?

4. Rise of Autonomous Agents

The possibility of autonomous AI agents acting independently is one of the next big frontiers for OpenAI. Such agents could handle workflows and decision-making without human intervention, marking a significant leap in how AI integrates into our lives. Combined with advances in consumer robotics, this could be a game-changer. Yet, the concept of autonomous agents also brings concerns—what does autonomy mean for accountability? If an AI agent makes a mistake or causes harm, who is responsible? The prospect of giving AI more independence is exciting but requires a robust framework for ethical oversight.

5. Innovation vs. Safety

Balancing innovation with safety remains crucial for OpenAI. They have taken steps to ensure iterative, safe releases while facing the challenge of progressing in a responsible manner. However, there have been concerns about departures from the safety-focused teams, raising questions about the emphasis on safety versus rapid product release. Is OpenAI truly balancing safety and innovation, or is safety taking a backseat to the competitive race for dominance in AI technology? The recent high-profile departures suggest that maintaining a culture of safety might be more challenging than initially presented. What did Ilya see?

6. Transforming Productivity

Recent updates suggest a future where AI becomes an active player in daily productivity. Instead of passively answering questions, AI could soon support, plan, and execute tasks autonomously, streamlining both personal and professional workflows. This level of integration could be transformative, but it also raises issues of dependency—how much autonomy should we be comfortable giving AI in our work and personal lives? Could over-reliance on AI for productivity eventually diminish our problem-solving skills?

What people cared about the most

• A folder for organising chats into research topics and projects in ChatGPT

• Advanced voice loosening restrictions around musical capabilities like singing

• Using ChatGPT as a therapist

• Increased ChatGPT memory per user and larger context windows

• Whether Sam was using ChatGPT to answer the questions to which he replied "sometimes, yes...can you tell?"

• If ChatGPT will be able to perform tasks on its own like messaging the user to which the answer is yes and it'll be big theme in 2025

The rabbit hole 🐰

• _{Follow the actual AMA thread here} 🔗

• _{Q&A’s that might have gone unnoticed} 🔗

• _{Can we continue to scale AI?} 🔗

• _{Safety not first at OpenAI says former researcher} 🔗

💙 Other tech news we’re loving

• Anthropic's Desktop Apps for Claude: Anthropic releases new desktop apps for Windows and Mac, making it easier to access their advanced AI assistant, Claude, on any device.

• Microsoft Delays Recall Feature Yet Again: Microsoft has delayed its Recall feature once more, now targeting December for testing via the Windows Insider program.

• Google Maps Introduces New Gemini AI Features: Google Maps unveils enhanced features powered by Gemini AI, allowing users to ask real-time questions about locations and receive AI-driven answers.

• Boston Dynamics’ Atlas Robot Gets Upgraded: Boston Dynamics reveals a new video showcasing Atlas’s improved humanoid capabilities, highlighting autonomy and adaptability.

• Claude’s New Analysis Tool with JavaScript Capabilities: Claude introduces a new tool enabling its AI to write and execute JavaScript code, pushing boundaries in interactive analysis.

• Meta Developing its Own AI-Powered Search Engine: Meta aims to reduce its dependence on major search providers with a proprietary AI search engine in the works.

• Google’s Project Jarvis to Automate Device Use: Google is building Project Jarvis, an AI designed to control computers autonomously, setting up a futuristic interface experience.

• Meta’s Notebook LM Competitor, Spirit LM: Meta introduces Spirit LM, an open-source model supporting both text and speech, rivaling Google’s Notebook LM in versatility.

What’s new: Projects

Meta’s LLaMA AI Powers U.S. Government Transformation

Meta’s Q3 2024 earnings call brought some game-changing updates for AI enthusiasts, as reported by The Verge. CEO Mark Zuckerberg revealed that Meta’s open-source AI model, LLaMA, is being rolled out to power various U.S. government initiatives. Collaborations with the U.S. State Department and Department of Education aim to leverage LLaMA to improve clean water access, infrastructure reliability, and even streamline student financial aid processes. The Q3 results also showed a strong performance, with $40.6 billion in revenue and $15.7 billion in net income, demonstrating the impact of LLaMA’s open-source model on reducing internal costs and improving content moderation.

Why This Matters to Me

For the AI community, Meta’s open-source approach with LLaMA represents a transformative shift. The model’s use in government projects underscores its versatility, potentially positioning LLaMA as a key player in the AI space, beyond the tech industry. This open-source strategy also signals Meta’s intention to compete with other AI giants like OpenAI and Anthropic, focusing on collaborative transparency to drive LLaMA’s adoption in various sectors

Additionally, Meta’s Q3 report shows a renewed commitment to AI, promising a "significant capital expenditure" increase in 2025 that will likely fuel further developments in LLaMA’s capabilities and accessibility

Rabbit Hole

Peering further into Meta’s strategy, we find an ambition to intertwine AI with governmental problem-solving. Imagine LLaMA improving resource management or streamlining bureaucratic processes—such moves could reshape how AI interacts with public infrastructure. Zuckerberg’s plans for an upcoming LLaMA update promise even greater reasoning capabilities and faster performance, hinting at breakthroughs on the horizon. As Meta navigates political scrutiny and increasing competition, this partnership with the U.S. government could set a precedent for AI’s role in public administration

For the latest on Meta’s ambitious AI future, read the full report on The Verge.

💙 Other Project news we’re loving

• The Government Data Quality Framework: The UK government unveils its Data Quality Framework to ensure accuracy and trustworthiness in government data reporting and management.

• BuiltWorlds Infrastructure Conference: At the 2024 Infrastructure Conference, experts discuss digital project management's transformative impact on infrastructure.

• Salesforce Releases AgentForce: Salesforce's AgentForce AI chatbot is now live, promising smarter customer service solutions powered by robust AI capabilities.

• LLMs to Analyse Complex Construction Contracts: New research explores LLMs in construction, helping automate contract analysis for improved project management and efficiency.

• Universal Music Group and Ethical AI Music Generation: Universal Music's partnership with Klay promises an ethically driven AI music generator, reshaping AI’s role in the creative arts.

• Costa Rica's AI-driven Brand Strategy: Costa Rica becomes the first country to use AI to create its national brand strategy, marking a new era in national identity development through technology.
  

What’s new: Productivity

Apple Intelligence first takes?

Summary

Apple is making waves in AI with its latest technology, "Apple Intelligence," now available on iPhone, iPad, and Mac. This new capability, introduced through an official announcement, promises to revolutionise user interactions with devices. Apple Intelligence focuses on personalised user experiences, including predictive text, smart suggestions, and more responsive voice interactions, all while prioritising privacy—a hallmark of Apple’s AI approach. This exciting update was also previewed in an informative YouTube video, which demonstrates how Apple’s latest features will change the way users interact with their devices.

Why This Matters to Me

For AI enthusiasts, Apple’s move underscores a broader trend of integrating generative and conversational AI into everyday tech products. Apple Intelligence aims to keep up with other tech giants in the AI race, blending advanced machine learning with Apple’s famously strong privacy features. As Apple embeds AI deeply within its ecosystem, this development could set a precedent for other consumer electronics, urging competitors to create more private, responsive AI experiences in their devices.

Rabbit Hole

Apple’s AI journey is worth watching for its unique approach—focusing on enhancing device functionality without compromising user privacy. Apple’s reliance on on-device processing, rather than cloud-based AI, aligns with its promise to protect personal data. This is in contrast to competitors like Google and Meta, which frequently leverage cloud processing for their generative AI. Apple’s AI could change how users think about privacy in an AI-driven world, pushing the industry to consider alternatives to cloud dependency. For a deeper look at how Apple Intelligence might transform user experiences, check out Apple’s full statement in their Newsroom post and the YouTube demo.

💙 Other productivity news we’re loving

• Udio’s New Track Editing Features: Udio rolls out improved track editing features, making music production and customisation faster and more intuitive for creators.

• AI Adoption Among Leaders: The 2024 AI Adoption Report from Wharton examines how leaders are embracing AI, highlighting trends, challenges, and industry applications shaping the future of human-technology interaction.

• AI-Replicated Michael Parkinson Hosts Unscripted Podcast: AI brings the late Michael Parkinson back to the mic in a celebrity podcast, showcasing cutting-edge conversational AI with an unscripted, lifelike experience.
  

Events

The AI Conference on November 14, 2024, is a must-attend virtual event for leaders exploring AI's return on investment. Top executives, like Scott Galloway and Bhavesh Dayalji, will share candid insights into the effectiveness of enterprise AI applications, discussing real-world cases, AI business models, and evaluating AI's actual business returns versus costs.

One More Thing

Hinton Warns we are making human intelligence irrelevant

Meet Project Flux: About Us

At Project Flux, we're committed to pioneering the future of construction and project delivery through the lens of cutting-edge Artificial Intelligence insights. Our vision is to be at the forefront of integrating AI into the fabric of project delivery, transforming how projects are conceptualised, planned, and executed.

Our Mission Project Flux aims to not only inform and educate but also to inspire professionals in the construction industry to embrace the transformative potential of AI. We believe in the power of AI to revolutionise project delivery, making it more efficient, predictive, and adaptable to the dynamic demands of the modern world.

What We Offer Through our insightful newsletters, podcasts and curated content on LinkedIn, and engaging discussions, Project Flux serves as a resource for professionals seeking to stay ahead in their field. We offer a blend of practical advice, thought leadership, and the latest developments in AI and construction technology.

What People Say…