Quora’s Poe Chatbot Lets Users Download Paywalled Articles Anytime

Amazon is bolstering its AI capabilities by hiring key executives from Adept, a San Francisco-based startup known for developing "agents" that automate enterprise workflows. This significant move demonstrates Amazon’s strategic focus on integrating advanced AI technologies to enhance operational efficiency and streamline processes.

David Luan, the co-founder and CEO of Adept, who also served as the vice president of engineering at OpenAI, will join Amazon. Alongside Luan, other co-founders of Adept, including Augustus Odena, Maxwell Nye, Erich Elsen, and Kelsey Szot, along with several other employees, will also transition to Amazon. Despite this leadership shift, Adept will continue to function as an independent entity with its remaining staff, maintaining its innovation and development trajectory.

A key aspect of this transition is Amazon’s use of some of Adept’s technology through a non-exclusive license. This arrangement allows Amazon to integrate Adept’s advanced technology into its operations while allowing Adept to continue its independent path and pursue further innovation.

David Luan will report directly to Rohit Prasad, a seasoned Amazon executive who leads the Artificial General Intelligence (AGI) team focused on developing large language model technology at Amazon’s Seattle headquarters. Prasad emphasized the importance of Luan and his team’s expertise in training state-of-the-art multimodal foundational models and building practical digital agents. He noted that this expertise aligns with Amazon’s vision to provide practical AI solutions to both consumer and enterprise customers. In a memo to employees, Prasad stated that the technology license from Adept would accelerate Amazon’s roadmap for building digital agents capable of automating complex software workflows.

Adept, which raised $350 million in a Series B funding round in March 2023, was valued at approximately $1 billion. The startup’s software is designed to help companies automate mundane tasks such as extracting information from documents, sending emails, processing applications, and more. These automation capabilities are crucial for businesses seeking to enhance efficiency and reduce manual workloads. The company has reportedly been in discussions with other tech giants, including Meta and Microsoft, about potential deals, reflecting its significant value and technological prowess.

The recruitment of Adept’s leaders is part of a broader trend where major tech corporations seek to partner with or acquire innovative startups to bolster their AI infrastructure and services. This trend is driven by the growing demand for AI-driven solutions and the competitive race to lead in AI technology development. However, AI startups face challenges, including high computing and labor costs, without substantial revenue streams, making strategic partnerships and acquisitions appealing.

Amazon’s deal with Adept mirrors a similar move by Microsoft, which recently hired Mustafa Suleyman, co-founder and former CEO of consumer chatbot startup Inflection AI, along with Inflection co-founder Karén Simonyan and other employees. These strategic hires reflect the tech giants’ ambitions to strengthen their AI capabilities through the acquisition of top talent and advanced technology from promising startups.

Regulatory scrutiny is increasing on AI deals between large tech corporations and smaller startups. The Wall Street Journal reported that the Federal Trade Commission (FTC) is investigating whether Microsoft structured a deal to gain control of Inflection AI without undergoing an FTC review. Microsoft had previously invested in Inflection AI. Additionally, in January, the FTC launched a separate inquiry into Microsoft’s investment in OpenAI and Amazon’s investment in Anthropic, another prominent AI startup.

These investigations highlight the regulatory challenges and scrutiny that accompany major tech acquisitions and investments, particularly in the rapidly evolving and highly competitive AI sector. Regulators are keen to ensure that these deals do not stifle competition or lead to monopolistic practices, thereby maintaining a fair and dynamic market environment.

Poe, an AI chatbot platform owned by Quora and backed by a $75 million investment from Andreessen Horowitz, has recently raised eyebrows by offering downloadable HTML files of articles from paywalled journalistic outlets. By prompting Poe’s Assistant bot with the URL of a paywalled article, users receive a detailed summary and a downloadable HTML capture of the entire article. This capability extends to major publications like The New York Times, Bloomberg Businessweek, The Atlantic, Forbes, Defector, and 404 Media.

This practice has sparked significant controversy, with critics arguing that it constitutes a blatant violation of intellectual property law. James Grimmelmann, a professor of digital and information law at Cornell University, labeled the practice as prima facie copyright infringement, as it involves making copies on Quora’s servers. Quora disputes this claim, comparing Poe to cloud storage services that store content at the user’s direction.

Further scrutiny reveals that when the Assistant bot is prompted to summarize a website, a server identifying itself as “Quora Bot” visits the site, often ignoring the Robots Exclusion Protocol—a web standard intended to guide automated access to websites. This approach has led to legal and ethical concerns from media executives who have observed similar patterns of access to their sites following prompts to Poe’s chatbot.

Quora’s response emphasizes that Poe’s file attachments are created at the direction of users and are consistent with other services that store and clip web content. However, media companies argue that the practice undermines their business models and violates copyright laws. For instance, The New York Times has explicitly prohibited scraping or reproducing its content without prior permission, highlighting the legal complexities of Poe’s approach.

Quora’s cofounder Adam D’Angelo, a former Facebook CTO and OpenAI board member, envisions a future where Poe and Quora are closely integrated, sharing datasets and AI capabilities. This vision aims to blend Quora’s human-centric knowledge base with Poe’s advanced AI-driven interactions, creating a more comprehensive and intelligent platform. Poe offers various chatbots, including those powered by Anthropic’s Claude model, designed to handle text provided by Quora’s crawlers. This capability allows users to retrieve and download full-text articles in HTML format, effectively bypassing paywalls and creating significant copyright concerns.

As tech giants like Amazon and Microsoft continue to expand their AI capabilities through strategic hires and acquisitions, the landscape of AI development is becoming increasingly complex and competitive. Simultaneously, platforms like Poe push the boundaries of content accessibility, raising important questions about intellectual property and the future of digital content distribution. Regulatory bodies and media companies are closely watching these developments, seeking to balance innovation with legal and ethical considerations.