Why Hollywood's AI Video Pivot Is Stumbling Over Prompts?

Here is what’s new in the AI world.

Directors vs. AI: Hollywood's New Creative Battle Over Video Prompts

The internet is now flooded with low-quality AI-generated videos, from bizarre animal antics to fake disasters that somehow garner thousands of enthusiastic comments praising them as the future of filmmaking.

While most outputs remain visually jarring, a growing contingent believes this technology poses an existential threat to traditional Hollywood production.

Enter Asteria, a new production studio founded by filmmaker Bryn Mooser that's taking a radically different approach to AI video generation.

Unlike typical AI video tools that produce uncanny results from simple text prompts, Asteria has developed an "ethical" AI model called Marey that:

Mooser argues that most AI video tools fail filmmakers because:

1. They're built by tech teams without filmmaking experience

2. The "text-to-blockbuster" fantasy ignores creative realities

3. Visual control remains too limited for professional work

"Christopher Nolan can't just text his way to 'The Odyssey'," Mooser notes, highlighting the gap between Silicon Valley promises and filmmakers' needs.

Asteria's Production Model

The studio's workflow was demonstrated in the animated short "A Love Letter to LA":

• Trained on 60 original illustrations by artist Paul Flores

• Generated consistent 2D assets converted to 3D environments

• Maintained artistic control while leveraging AI efficiency

The company is pioneering new ownership models where:

• Filmmakers retain partial rights to their custom AI models

• Revenue-sharing options are being explored

• Production costs could drop from $150M to $10-20M range

The Ultimate Test: "Uncanny Valley"

Natasha Lyonne's upcoming feature will serve as Asteria's proving ground:

• Live-action story about reality distortion

• AI-generated "video game" elements as an intentional aesthetic

• Designed to make the technology invisible to audiences

Mooser envisions AI as a democratizing tool rather than a replacement: "Nobody wants to just see what computers dream up. The human touch remains essential."

As Hollywood grapples with AI's implications, Asteria represents one vision for how the technology might integrate into filmmaking, not as a magic button, but as another brush in the artist's toolkit. The coming years will reveal whether this approach can deliver on both creative and commercial potential.

VS Code Takes First Steps Toward Open-Source AI Integration

In a significant move toward open AI development ecosystems, Microsoft has released the GitHub Copilot Chat extension's source code under the MIT license, marking the first step in transforming Visual Studio Code into a fully open-source AI editor.

The June 30 announcement signals Microsoft's commitment to transparent, developer-controlled AI tooling.

Key Technical Revelations

The open-sourced codebase exposes previously proprietary elements:

• Agent mode implementation

• Context engineering approaches

• Telemetry collection practices

• Complete system prompts the architecture

Strategic Implications

1. Developer Trust Building - "Every implementation detail is now visible," states the VS Code team, addressing longstanding concerns about opaque AI systems

2. Community Integration - Microsoft actively seeks contributor input with plans to merge the extension into VS Code's core repository

3. Market Positioning - Erich Gamma emphasizes that this responds to enterprise demand for open-source IDE alternatives

The decision follows:

• Satya Nadella's Build 2025 commitment to embedded AI development

• Growing standardization of LLM-powered coding interfaces

• Competitive pressure from VC-backed alternatives like Cursor (valued at $500M ARR)

While maintaining proprietary control over inline completions for now, Microsoft plans to:

• Open-source prompt testing infrastructure

• Gradually incorporate all Copilot functionality

• Maintain MIT-licensed transparency for core chat features

The move has sparked industry debate, with some noting the irony of Microsoft's open approach compared to OpenAI's acquisition of closed competitor Windsurf for $3B.

As AI development tools mature, Microsoft appears to be betting that transparency and extensibility will drive wider adoption than walled-garden alternatives.

After DeepSeek's Success, Baidu Counters With Open-Source Ernie

China's tech giant Baidu has announced plans to open-source its Ernie generative AI model, marking a significant strategic shift for the company and potentially disrupting global AI economics.

The phased release represents China's most substantial open-source AI move since DeepSeek's emergence, though analysts debate its market impact.

Strategic Reversal with Global Implications

Baidu's decision reverses its longstanding proprietary approach:

• Previously opposed open-source models as inferior to proprietary solutions

• Now acknowledges the competitive viability of open-source AI (citing DeepSeek's success)

• Could pressure closed-model providers like OpenAI and Anthropic to justify premium pricing

Market Disruption Potential

Industry experts highlight several potential impacts:

1. Price Wars: Baidu claims Ernie X1 matches DeepSeek R1 performance at half the cost

2. Developer Empowerment: Lowers barriers for global AI application development

3. Geopolitical Tensions: Raise security concerns about Chinese AI in Western markets

Mixed Industry Reactions

OpenAI's Dilemma

Sam Altman has acknowledged the need for OpenAI to reconsider its open-source strategy:

• Plans the summer 2023 release of an open model (since delayed)

• Admits that maintaining a technological lead is becoming harder

• Faces pressure to balance proprietary interests with developer demands

Ongoing Challenges

Even with open-sourcing, key issues remain unresolved:

• Training data transparency and compensation

• Enterprise security concerns

• Geopolitical tensions affecting adoption

• Customization requirements for practical implementation

As the AI industry matures, Baidu's move accelerates the open vs. closed debate while testing global markets' willingness to embrace Chinese AI technology at scale. The coming months will reveal whether this represents true disruption or merely another option in an increasingly crowded field.

Mark Chen Warns Staff: OpenAI Pauses All Work for 7 Days, Meta Knows

OpenAI is instituting an unusual week-long company closure next week to help employees recover from months of intensive 80-hour workweeks, according to Wired sources.

The temporary shutdown comes as the AI pioneer battles to retain top researchers against Meta's aggressive recruitment campaign, offering packages reportedly worth up to $100 million.

Key Details of the Shutdown:

• Only executive staff will continue working during the break

• Affects all non-critical operations at the ChatGPT developer

• Follows sustained pressure to advance artificial general intelligence (AGI) development

Meta's Recruitment Onslaught

OpenAI's Chief Research Officer Mark Chen warned employees in an internal Slack message that Meta would likely intensify its poaching efforts during this vulnerable period.

The social media giant has already successfully recruited seven OpenAI researchers in recent weeks, including:

• Lucas Beyer

• Alexander Kolesnikov

• Xiaohua Zhai

• Trapit Bansal (key contributor to OpenAI's o1 model)

Compensation Reassessment

In response to the talent drain, OpenAI leadership is:

• Recalibrating compensation packages

• Developing "creative" reward systems for top performers

• Shifting focus from frequent product releases to AGI breakthroughs

Leadership's Stance

CEO Sam Altman has dismissed reports of $100M+ Meta offers as "fake news," while maintaining that OpenAI's mission and innovation culture provide superior retention value.

The shutdown represents a strategic pause for the normally relentless AI developer as it navigates unprecedented talent competition in the red-hot artificial intelligence sector. Industry analysts suggest this may mark a turning point in how Silicon Valley's AI labs balance aggressive development timelines with employee sustainability.

Jony Ive and OpenAI's First AI Device: A Pen-Like Gadget, Report Says

The AI leader is developing its inaugural hardware product – reportedly a context-aware smart pen marking its first major foray into consumer devices. This follows OpenAI's strategic merger with Jony Ive's LoveFrom design firm in May, which hinted at creating "a family of AI-native devices."

Key Features (Based on Reports & Filings):

Strategic Context:

• Represents OpenAI's first hardware product since ChatGPT's launch

• Combining Ive's iconic Apple design pedigree with OpenAI's AI capabilities

• Aims to create a new $1T market opportunity (per Altman's internal comments)

• Differentiates from smartphone-centric AI approaches

  The device could:

1. Establish a new hardware category for ambient AI assistance

2. Leverage OpenAI's language models in physical contexts

3. Compete indirectly with smartphone-based AI assistants

4. Showcase practical applications of multimodal AI

While details remain speculative, the project signals OpenAI's ambition to move beyond software and shape how consumers interact with AI in physical spaces. The pen form factor suggests a focus on creative and productivity use cases, potentially revolutionizing how professionals interface with generative AI.

Your opinion matters!

Hope you loved reading our piece of newsletter as much as we had fun writing it. 

Share your experience and feedback with us below ‘cause we take your critique very critically. 

Thank you for reading

-Shen & Towards AGI team