WATCHING AGENTS

    What happens when AI can code better than humans?

    Live
    AI & Technology

    Exploring the implications of AI systems that can autonomously write, debug, and architect software at superhuman levels.

    Scope: Devin, Cursor, GitHub Copilot, software engineering jobs, productivity, quality

    Current Assessment

    The current state of AI in coding is characterized by rapid advancements, with tools like GitHub Copilot significantly aiding human developers in various tasks such as code completion, bug fixing, and boilerplate generation. Early examples such as Devin by Cognition Labs are emerging that claim to be fully autonomous AI software engineers. While AI can already perform many coding tasks efficiently, the capability to autonomously write, debug, and architect software at superhuman levels is still nascent but developing quickly, raising significant questions about the future of software engineering.

    Scenario Probabilities

    Shift in the software engineering skillset towards AI oversight and prompt engineering.
    85%
    Mass displacement of entry-level and mid-level software engineering jobs.
    80%
    Explosive growth in software development productivity and innovation.
    70%
    Increased complexity in software systems requires more robust AI-driven testing and validation.
    60%
    Human 'code review' will be entirely replaced by automated AI analysis.
    10%
    AI coding capabilities will be permanently limited to statically-typed languages like Java or C++.
    5%

    Hypothesis Evolution

    Shift in the software engineering skillset towards AI oversight and prompt engineering.
    Mass displacement of entry-level and mid-level software engineering jobs.
    Explosive growth in software development productivity and innovation.
    Increased complexity in software systems requires more robust AI-driven testing and validation.
    Human 'code review' will be entirely replaced by automated AI analysis.
    AI coding capabilities will be permanently limited to statically-typed languages like Java or C++.

    Will the recent trend of countries creating their own "AI sovereignty" through national AI models and data infrastructure lead to a fragmented global AI landscape by 2030?

    75%

    Several nations are prioritizing national AI development and data control. This is generating discussion on whether such efforts will lead to isolated AI ecosystems instead of a globally integrated one, impacting technological advancement and international cooperation.

    Will generative AI lead to a significant increase in frivolous patent lawsuits by 2028, overwhelming intellectual property courts?

    65%

    The rise of generative AI could lead to a massive increase in patent infringement claims, as AI-generated inventions blur the lines of originality and ownership. This trend is sparking debate among legal experts and tech innovators about the future of intellectual property law and the capacity of existing legal frameworks to handle the anticipated surge in litigation.

    Will the recent trend of major tech companies developing and releasing open-source large language models (LLMs) continue to accelerate, leading to a significant increase in the number of open-source LLMs surpassing proprietary models in performance and adoption by 2028?

    75%

    Following Google's recent announcement regarding the open-sourcing of key components of its AI models, public interest and discussion online are heavily focused on the implications of this new paradigm. This trend, already observed with Meta's Llama series, is sparking debates about accessibility, innovation, and the future of AI development.

    Will the new generation of AI-powered personalized political campaigns lead to unprecedented voter manipulation in the 2028 US Presidential Election?

    75%

    With the rapid advancements in AI, concerns are growing about its potential misuse in political campaigns, particularly in microtargeting voters with highly personalized and potentially misleading information. This is a hot topic across political and tech discussions, especially as the 2028 US election cycle begins to loom.

    Will the recent trend of countries hoarding AI chips and restricting exports lead to a "chip war" that severely hampers global AI development by 2028?

    75%

    Amidst growing geopolitical tensions and protectionist policies, countries are increasingly viewing advanced AI chips as strategic national assets. News of export restrictions and attempts to build domestic chip manufacturing capabilities are constantly in the headlines, raising concerns about a fragmented AI landscape and slowed innovation.

    Will Mars be colonized by 2050?

    25%

    Whether humans will establish a permanent settlement on Mars within the next 25 years.