• Project Flux
  • Posts
  • Google's Gemini 3 Pro Launches and AI Image Tools Cross a New Threshold

Google's Gemini 3 Pro Launches and AI Image Tools Cross a New Threshold

From Google's muted model release to Parliament's billion pound data wake up call, discover what actually matters for project delivery this week.

Proudly sponsored by ConstructAI, brought to you by Weston Analytics.

Hello Project AI enthusiasts,

Another week, another wave of AI announcements competing for your attention. Here at Project Flux, we cut through the noise to bring you what actually matters for project delivery. This week saw Google launch Gemini 3 Pro to surprisingly muted applause, Parliament finally recognise that we're burning billions in project intelligence, and LinkedIn get caught with its algorithmic bias exposed. Meanwhile, Bezos decided retirement was boring and designers started updating their CVs.

In This Edition

Flux check-in

Google's Gemini 3 Pro Arrives But the AI Wow Factor Is Fading

Google unveiled Gemini 3 Pro this week, and whilst the response has been more measured than previous releases, this signals an encouraging maturity in the field. We're moving beyond the "more compute equals better AI" era into a new phase focused on practical application. For project delivery teams, this is welcome news, as the real gains now come from integration rather than waiting for the next big model innovation. The foundation models have reached impressive capabilities, and now we're entering the valuable phase of putting them to work effectively. Read the full breakdown →

What Does This Mean for Me?

For project delivery professionals, this signals a fundamental shift in AI strategy. Stop chasing every new model release and start focusing on practical implementations that integrate with your existing systems. The real competitive advantage now lies in how well you connect AI to your project management software, your industry terminology, and your team's workflows. A slightly less capable model that plays nicely with Microsoft Project beats a genius model that lives in isolation. Focus your energy on integration and adoption, not benchmarks.

Key Themes

  • Diminishing Returns: Scaling returns are diminishing as each improvement requires exponentially more resources and billion dollar training runs

  • Multimodal Processing: Better multimodal processing now handles architectural drawings, Gantt charts and project documentation together seamlessly

  • Integration Innovation: Real innovation is happening in middleware and integration layers where practical value actually lives for project teams

  • Practical Focus: Stop chasing every model release and start building practical implementations that integrate with your existing workflows

Down the Rabbit Hole

Why Westminster's Data Awakening Matters for Your Projects

MPs have declared project data a national asset, shocking nobody who has worked on major UK projects. Martin Paver from Projecting Success has made this case for years, noting we let so much data go up in a plume of smoke. The National Audit Office found 67% of project failures stem from known issues previous projects encountered. We are paying billions to rediscover solved problems. Read the full breakdown →

What Does This Mean for Me?

Project managers drowning in repeated mistakes finally have political backing for change. But recognition without action is just expensive noise. If you work on major programmes, start documenting your project data systematically now. Build the case internally for knowledge retention before your current project ends. The organisations that get ahead of this will be those who treat project intelligence as an asset to be preserved, not waste to be discarded. Push for data contribution requirements in your contracts and procurement processes.

Key Themes

  • Data Destruction: UK destroys terabytes of invaluable project intelligence when each programme ends costing the nation billions annually

  • Repeated Waste: £15 billion annual cost from poor project delivery represents pure waste from mistakes we keep repeating unnecessarily

  • Knowledge Loss: Only 8% of government projects systematically capture and share lessons learned leaving 92% knowledge to vanish

  • Political Barriers: Technical barriers account for just 15% of data sharing failures with organisational politics being the real blocker

Down the Rabbit Hole

Bezos Returns to Build the AI Infrastructure Everyone Will Need

When Jeff Bezos stepped down from Amazon in 2021, nobody believed the retirement story. This week's announcement of Project Prometheus, a $6.2 billion AI venture where he serves as co-CEO, proves doubters right. Is this ego to keep up with Musk? Bezos understands complex operations better than almost anyone alive, and his boring infrastructure approach could reshape AI tools for project delivery entirely. Read the full breakdown →

What Does This Mean for Me?

When Bezos enters a market seriously, the industry changes whether you like it or not. For project delivery teams, expect infrastructure plays that others will depend on and potentially more affordable AI tools through his scale-driven cost reduction model. Watch for Prometheus to focus on enterprise solutions where the real money lives, not consumer chatbots. If history repeats, the tools emerging from this venture in two to three years could become as essential to AI workflows as AWS became to cloud computing. Position yourself to benefit from the tools they create while avoiding becoming collateral damage in their competition.

Key Themes

  • Serious Commitment: Bezos returning as hands-on co-CEO signals serious long-term commitment not just passive investment from the sidelines

  • Infrastructure Play: Classic Bezos strategy focuses on owning the infrastructure layer that everyone else will eventually need to depend on

  • Talent Acquisition: Already hired over 100 engineers from Google, Meta and OpenAI with compensation packages that dwarf standard tech salaries

  • Quiet Execution: Expect quiet execution for two to three years followed by emergence with something boringly essential like AWS was

Down the Rabbit Hole

Google's AI Image Tools Cross the Professional Threshold

Google launched Nano Banana Pro this week with less fanfare than a parish council meeting, but implications are seismic. The integration with Adobe Firefly and Photoshop, combined with Notebook LM's new capabilities, represents the moment AI image generation became professionally useful. Remember AI art from 2022 with melting faces? Error rates dropped 94% since then. This is now a professional tool. Read the full breakdown →

What Does This Mean for Me?

Project documentation is about to get aggressively visual, and that 200-page requirements document nobody reads is about to look as dated as a fax machine. For project managers, this means you can finally communicate complex information in formats stakeholders will actually engage with. Start experimenting with NotebookLM for your status reports and risk registers now. The barrier to professional visual communication has effectively disappeared. Those who master visual project storytelling will have a significant advantage in stakeholder management and decision-making speed.

Key Themes

  • Instant Infographics: Notebook LM can now convert 50-page reports into single-page infographics automatically saving hours of design work

  • Speed Revolution: Human designers spend four to six hours on infographics while Nano Banana Pro delivers professional results in seconds

  • Seamless Adoption: Adobe partnership solves enterprise adoption headaches because there is no new software to learn or procurement battles

  • Better Comprehension: Visual documentation improves stakeholder comprehension by 74% according to PMI research on project communications

Down the Rabbit Hole

LinkedIn's Algorithm Bias Against Women Finally Exposed

A grassroots investigation on LinkedIn revealed what many suspected but could not prove: the platform's algorithm systematically favours content from male profiles. Women changing their profile gender settings to male report dramatic engagement increases. Posts from male profiles receive 2.3 times more views on average. LinkedIn's response has been textbook Silicon Valley crisis management: acknowledge, promise investigation, change nothing. Read the full breakdown →

What Does This Mean for Me?

If you are a female project professional, this visibility gap could be costing you £12,000 to £15,000 annually in lost opportunities, missed connections, and suppressed thought leadership reach. LinkedIn has become critical infrastructure for professional development and career advancement. Consider the experiments yourself, document any discrepancies, and join the growing movement demanding algorithmic transparency. For organisations, this is a reminder that the platforms you rely on for talent acquisition and professional networking may be systematically disadvantaging half your potential workforce. Factor this bias into your recruitment and visibility strategies.

Key Themes

The pulse check

Tips of the week

MCP Servers: Give Your AI Superpowers Beyond Just Chat

Imagine Claude or ChatGPT as a brilliant consultant locked in a room. They can answer questions, but they cannot check your emails, search your files, or update your CRM. MCP (Model Context Protocol) servers are the keys that unlock that room. They are connectors that let AI tools directly interact with your software like Outlook, Google Drive, and Slack. Instead of copying and pasting information back and forth, the AI can now reach into those systems itself.

Both Claude and ChatGPT offer native integrations you can enable in settings with no technical setup. Beyond built-in options, explore mcpservers.org for community-built servers covering Notion, databases, and more. Once you enable the Filesystem MCP, Claude can even install and configure other MCP servers for you.

The initial setup takes 10 minutes. The time saved is hours every week. MCP servers turn AI from a tool you use into an assistant that works for you.

Need help structuring your prompts? Try my free Prompt Generator. Describe your task, select parameters (intent, tone, length), and get a structured prompt ready to paste into ChatGPT, Claude, or Perplexity in 30 seconds.

If you would like a deeper, structured learning path, here is the Referral Link for the Course

 

Governance & Security

The regulatory landscape shifts rapidly as governments grapple with AI's accelerating capabilities. The EU's GDPR enforcement faces new challenges in the AI era, with regulators questioning whether existing frameworks adequately address algorithmic decision-making at scale. Meanwhile, President Trump warned against AI overregulation, arguing the US needs a single federal standard rather than a patchwork of state rules, significantly impacting how American AI companies operate globally.

The tension between innovation and oversight has never been more acute. Reports indicate enforcement mechanisms struggle to keep pace with technological advancement. For project delivery professionals, this creates both opportunity and risk: clearer regulatory frameworks could simplify compliance, but current uncertainty makes long-term planning challenging. The coming months will see significant movement on both sides of the Atlantic. EU regulators push for stricter enforcement while US policy favours lighter-touch approaches. Organisations operating internationally must navigate these diverging philosophies carefully, building flexibility into compliance strategies.

The European Commission has put forward plans to simplify and relax several technology regulations, including pushing back certain parts of the AI Act, in an effort to reduce bureaucracy, address criticism from major tech companies, and strengthen Europe’s competitiveness. This approach follows similar adjustments to some environmental rules after objections from businesses and the United States government.

Robotics

UBTECH Humanoid Robot Walker UBTECH has begun mass production and delivery of the first batch of several hundred full-size industrial humanoid robots, Walker S2, which will be deployed in phases across frontline industrial applications. This milestone moves the company toward its target of delivering 500 units within the year and marks the beginning of large-scale, real-world implementation of humanoid robotics. See how it works

Waymo enters 3 more cities: Minneapolis, New Orleans, and TampaThe company said on Thursday that it will start manually driving its vehicles in those cities “in the coming days,” ahead of testing and validating its driverless tech, before eventually launching a commercial robotaxi service in each. Explore Waymo

NASA Deep Space Antenna Damage, Russian Robot Fail & Gross GPU RestoreNASA Deep Space Network damage, Russian AI robot fail, GPU restoration from smoker tar, AI training data chores, and space food from urine. Explore wild tech

AI platforms for defense applications Investors are putting more money into military robotics. Forterra last week raised $238 million in Series C equity and debt. The company said it plans to use the funds to support continued innovation in communications, command, and control systems. Forterra also plans to enhance production capacity for edge computing platforms serving defense and emerging mission domains.  Robot report

  • Payapps’ innovative payment technologyPayappas, an Autodesk company and one of the UK’s leading suppliers of automated payment software, saw off strong competition to be ‘Highly Commended’ in the prestigious Payment Technology award category at this year’s Global Fintech Awards. Read the full story 

  • Google unveils Gemini 3 AI model and AI-first IDE called Antigravity Google says the release of Gemini 3 is yet another step toward artificial general intelligence (AGI). The new version of Google’s flagship AI model has expanded simulated reasoning abilities and shows improved understanding of text, images, and video. Read the full update

  • Cursor Raises $2.3B Bringing It to a $29.3B Valuation The AI funding juggernaut shows no signs of slowing down, with coding development platform Cursor the latest to unveil massive new investment courtesy of a Series D round that raised $2.3 billion at $29.3 billion valuation. Read the full update

  • Flowith Captivates Korean Tech Community with Vision for AI Agent EcosystemAt this year's AI Summit Seoul, global attendees witnessed a new chapter in the human-AI relationship as Flowith, a leading AI agent creation workspace, introduced FlowithOS, the world's first operating system built natively for AI agents, solidifying its position as a key innovator in the rapidly evolving AI landscape. Dive into the full update

  • Intuit and OpenAI Join Forces to Revolutionize Financial IntelligenceFor the first time, consumers and businesses will be able to unlock the power of Intuit’s AI-driven expert platform and get personalized financial insights and recommendations they can seamlessly act on, all through apps available in ChatGPT. Explore the announcement

  • Adobe integrates Google Gemini 3 (Nano Banana Pro) into Firefly & PhotoshopAdobe has announced the integration of Google’s latest image model, Gemini 3 (with Nano Banana Pro), into its flagship creative applications, Adobe Firefly and Photoshop, effective immediately. Learn more

  • Microsoft’s new Anthropic partnership brings Claude AI models to AzureMicrosoft is announcing a strategic partnership with Anthropic today that will bring the AI startup’s models to Microsoft Foundry for the first time. As part of the deal, Anthropic is also committing to purchasing $30 billion of Azure compute capacity. Review technical details

  • ChatGPT launches group chats globallyChatGPT is launching group chats globally to all users on Free, Go, Plus, and Pro plans. Discover the details

Links We are Loving

Community

The Spotlight Podcast

Sean Betts and Paul Thorpe on When AI Meets Construction: Lessons in Embracing Uncertainty.

The latest Project Flux podcast, recorded live from an Overbury event, brings together Sean Betts, Chief AI and Innovation Officer at Omnicom Media Group UK, and Paul Thorpe, Digital Director at Okana. What happens when AI innovation meets construction's reality? A conversation exploring how skills and attitude are being levelled at warp speed, giving newcomers the same tools as industry veterans. Paul paints AI not as a role disruptor, but as the great leveller closing inefficiency chasms that have long-standing construction for decades. Key insight: you don't want the drill, you want the hole. Focus on outcomes, not tools. The future belongs to the adaptable.

Event of the Week

International Conference on Application of Artificial Intelligence and Internet of Things on Management, Science and Technology (ICAAIITMST)

2nd December 2025 | London, UK

International Conference on Application of Artificial Intelligence and Internet of Things on Management, Science and Technology is an organisation dedicated to supporting professionals across all fields and countries through strategically designed activities, including international conference organisation and journal publications. It also provides access to critical resources and aids global policymaking. It ultimately aims to facilitate learning, networking, and innovation for faculty, students, professionals, and policymakers from any domain. Register now


One more thing

An example of what Nano Banana Pro in NotebookLM can produce when you ask it to explain Gemini 3 Pro

That’s it for today!

Before you go we’d love to know what you thought of today's newsletter to help us improve The Project Flux experience for you.

Login or Subscribe to participate in polls.

See you soon,

James, Yoshi and Aaron—Project Flux 

1