E170: Tech's Vibe Shift, TikTok ban debate, Vertical AI boom, Florida bans lab-grown meat & more
On this episode of All-In with Chamath, Jason, Sacks & Friedberg, the hosts engage in a multifaceted discussion on several pressing issues in the tech landscape. They delve into the proposed TikTok ban and ByteDance's divestiture, exploring concerns surrounding national security risks and free speech implications. The conversation also touches on the use of copyrighted data for training AI models like ChatGPT, raising questions about derivative works and fair use.
Additionally, the hosts examine the rise of vertical AI startups tailored to specific professional roles, and the potential impact of AI on job automation. Sparking further debate is Florida's proposed ban on lab-grown meat, which highlights tensions between consumer choice, innovation, and regulatory protectionism for incumbent industries.
This is a preview of the Shortform summary of the Mar 16, 2024 episode of the All-In with Chamath, Jason, Sacks & Friedberg
Sign up for Shortform to access the whole episode summary along with additional materials like counterarguments and context.
1-Page Summary
TikTok Ban Bill and Divestiture of ByteDance Ownership
The U.S. House of Representatives has passed a bill that could either ban TikTok or force its sale by ByteDance, with President Biden indicating readiness to sign it into law. The bill aims to counter the control of applications by foreign entities like China over national security and foreign influence worries. Jason Calacanis stresses the control the Chinese government has over its citizens' access to apps, questioning the accessibility of a Chinese app like TikTok in America. Meanwhile, Chamath Palihapitiya points to the strategic restriction of U.S. services in China, advocating for balanced access. David Sacks proposes using trade bills to negotiate access between China and the U.S.
Some progressives voice concerns that the bill may suppress free speech, particularly on issues such as the Israel/Palestine conflict, due to TikTok's role as a platform for political expression among young users. National security concerns are further discussed regarding TikTok's data collection and surveillance capabilities, including anecdotal evidence that TikTok might access data through smartphone microphones. The speakers, including Palihapitiya and Calacanis, encourage software controls to prevent potential abuses, and they highlight the general agreement in Washington on the threat posed by foreign-controlled apps.
Opening AI Training Data and Derivative Works
Controversy surrounds OpenAI's potential use of unlicensed YouTube content for training ChatGPT. The CTO's refusal to specify data sources and the model's behavior raise questions about copyright and derivative works. The New York Times has sued OpenAI over derivative works and licensing issues, leading David Friedberg to argue that using open internet sources for AI training shouldn't be viewed as creating derivative works. David Sacks suggests that OpenAI might defend its practices under fair use doctrine, while Jason Calacanis discusses who should hold rights to derivative works, questioning if they should reside with original IP owners or be available to developers. Chamath Palihapitiya expresses concern about the legal expertise available to adjudicate complex copyright issues entangled with tech advancements.
Emergence of Vertical AI Startups
Vertical AI startups are becoming prominent for their focused development of AI applications for specific professional roles. Startups like Harvey, Abridge, TaxGPT, and Sierra develop tools tailored to legal, medical, tax, and customer support sectors. Cognition's Devon is highlighted for its exceptional coding competencies, possibly built upon GPT-4, suggesting it could rival human engineers.
Discussion shifts to the potential and limitations of these AI agents. While full autonomy might be years away, Friedberg envisions AI dramatically expanding human capability, while Palihapitiya predicts that future companies might be operated by individuals with AI-driven software. Calacanis observes a trend towards smaller teams or entrepreneurs leveraging AI. The specialized use of AI in certain fields is foreseen to eventually coordinate work across various roles, significantly supplementing or replacing human jobs.
Florida Banning Lab-Grown Meat
The Florida state legislature is considering a ban on lab-grown meat, sparking debates over the influence of incumbent ranchers and implications for consumer choice and innovation. The proposed bill suggests a classic case of regulatory capture, as it is perceived to protect incumbent meat producers’ interests. Friedberg criticizes the bill as limiting consumer choice and as an obstacle to technological progress. Drawing comparisons to past technological advancements, he argues that regulatory actions should not block new technologies that provide sustainable and ethical market alternatives. The conversation reflects concerns that regulations driven by industry lobbying could hinder consumer options and suppress technological innovations in food production.
1-Page Summary
Additional Materials
Clarifications
- ByteDance is a Chinese internet technology company known for developing popular apps like TikTok and Douyin. It has faced scrutiny in various countries over concerns related to security, surveillance, and censorship. Founded in 2012 by Zhang Yiming and Liang Rubo, ByteDance is headquartered in Beijing and incorporated in the Cayman Islands. The company also created the news platform Toutiao.
- ChatGPT is a chatbot developed by OpenAI that uses large language models to facilitate conversations with users. It allows users to guide the conversation's length, style, and level of detail. ChatGPT gained significant popularity after its launch and has both free and paid versions for users. It is based on OpenAI's generative pre-trained transformer (GPT) models and is designed for conversational applications.
- The fair use doctrine is a legal principle that allows the limited use of copyrighted material without permission from the copyright owner for purposes such as criticism, commentary, news reporting, teaching, scholarship, or research. It aims to balance the rights of copyright holders with the public interest in accessing and using creative works. Fair use is determined on a case-by-case basis by considering factors like the purpose of the use, the nature of the copyrighted work, the amount used, and the effect on the market for the original work. It provides flexibility in copyright law to promote creativity, innovation, and the free exchange of ideas.
- Derivative works in AI training involve using existing content to train AI models. This raises questions about copyright and ownership of the resulting AI model. Issues can arise when AI models are trained on copyrighted material without proper authorization. The legality of using such data for training AI models is a complex area that involves considerations of fair use and intellectual property rights.
- Regulatory capture is a concept where regulatory agencies, originally tasked with serving the public interest, end up being influenced or controlled by the industries they are supposed to regulate. This can lead to these agencies prioritizing the interests of specific groups or industries over the broader public good. The theory suggests that powerful interest groups can manipulate regulations to benefit themselves, often at the expense of the general population. Regulatory capture poses a risk to the effectiveness and fairness of regulatory bodies, potentially undermining their intended purpose.
- The debate in Florida regarding lab-grown meat involves discussions about potentially banning the production and sale of meat that is grown in a laboratory setting instead of being sourced from traditional livestock farming. This debate raises concerns about the influence of traditional meat producers, consumer choice, and the impact on innovation in the food industry. Supporters of lab-grown meat argue that it offers sustainable and ethical alternatives to conventional meat production methods, while critics suggest that regulatory actions like a ban could stifle technological progress and limit consumer options. The proposed ban is seen by some as an example of regulatory capture, where existing industry interests influence legislation to maintain their market dominance.
- Generative Pre-trained Transformer 4 (GPT-4) is a large language model developed by OpenAI, following the success of its predecessors like GPT-3. GPT-4 was designed to be more reliable, creative, and capable of handling nuanced instructions compared to earlier versions. It was launched in 2023 and introduced improvements such as larger context windows and the ability to process images alongside text inputs. GPT-4 has been used in products like ChatGPT Plus and Microsoft Copilot for various language-based tasks and interactions.
- AI agents have the potential to significantly enhance human capabilities by automating tasks and providing valuable insights. However, they currently have limitations in terms of full autonomy and understanding complex contexts like humans do. While AI can streamline processes and improve efficiency, it may not completely replace human roles in certain fields due to the nuanced decision-making and creativity required. The development of AI is expected to continue evolving, with a focus on augmenting human work rather than entirely replacing it.
- Industry lobbying influence on regulations involves businesses or trade organizations attempting to influence government policies and decisions in their favor. This influence can shape regulations to benefit specific industries, often at the expense of consumer interests or competition. Lobbying efforts can involve financial contributions, advocacy campaigns, and direct communication with policymakers to sway regulations in a way that aligns with the industry's goals. The impact of industry lobbying on regulations can lead to laws that prioritize the interests of certain businesses over broader societal concerns.
Counterarguments
- Concerns about TikTok's data privacy and national security may be overblown or not unique to TikTok, as many social media platforms collect extensive user data.
- Banning TikTok could be seen as a form of censorship and may infringe on the free speech rights of users who use the platform for expression and activism.
- The effectiveness of trade bills to negotiate app access between China and the U.S. might be limited due to the fundamentally different political and economic systems.
- The argument that using open internet sources for AI training doesn't create derivative works could undermine content creators' rights and the value of their original works.
- Fair use doctrine as a defense for AI training practices might not be applicable in all cases, especially if the AI's output closely resembles the original content.
- The rise of vertical AI startups could lead to over-reliance on technology in critical sectors, potentially causing ethical dilemmas and reducing the need for human expertise.
- The potential for AI to replace human jobs might exacerbate economic inequality and lead to a workforce that is not adequately prepared for a technology-centric future.
- Banning lab-grown meat could be seen as a necessary step to ensure food safety and protect public health until such products are proven to be safe and regulated appropriately.
- Regulatory actions against lab-grown meat might also be viewed as protecting cultural and traditional agricultural practices that are important to a region's identity.
Related Shortform Content
Get access to the context and additional materials
TikTok Ban Bill and Divestiture of ByteDance Ownership
The United States House of Representatives passed a bill that could result in a TikTok ban or force ByteDance to sell the app, which President Biden has signaled he would sign into law. This legislation targets applications controlled by foreign adversaries, such as China, and aims to address concerns of national security and foreign influence.
Reciprocity - no access to US apps and social platforms in China
The bill addresses the imbalance of access between U.S. and Chinese applications, noting that platforms like Instagram are not accessible in China. Calacanis emphasizes the Chinese government's control over their citizens and questions why the U.S. would allow a Chinese app such as TikTok to have considerable reach in America. Chamath Palihapitiya cites historical examples where China favors domestic brands and restricts access to U.S. products and services, showing strategic moves to control information and deny reciprocity.
David Sacks suggests that trade bills should be used to allow mutual access to products from China and the U.S. rather than a TikTok ban that grants expansive powers to the government.
Some fears of loss of free speech, particularly around Israel/Palestine issues
Progressives fear the bill could stifle political discourse, especially among young people. TikTok is viewed as a platform active in pro-Palestinian and anti-Israel discourse, leading to concerns that the bill could silence such discussions. There is also concern over the law's ambiguity and potential overreach, which could empower the President to act against companies supposedly aligned with foreign interests.
Potential security threat if CCP asserts control over users' feeds
TikTok, owned by ByteDance, could pose a security threat if the Chinese Communist Party (CCP) exercises control over users' feeds. The bill's broad language could impact any app considered to be under the influence of a foreign entity. Discussions included the app's potential to manipulate political issues and influence opinions.
Palihapitiya shares an anecdote suggesting TikTok can present ads based on private conversations, indicative of a surveillance capability that raises nat ...
Here’s what you’ll find in our full summary
TikTok Ban Bill and Divestiture of ByteDance Ownership
Additional Materials
Clarifications
- The bill passed by the United States House of Representatives targets applications controlled by foreign adversaries, such as China, with the aim of addressing concerns related to national security and foreign influence. It could potentially lead to a TikTok ban or force ByteDance to sell the app. The legislation seeks to create reciprocity in access between U.S. and Chinese applications, highlighting concerns about the Chinese government's control over its citizens and the imbalance in market access. Additionally, the bill raises fears about potential security threats if the Chinese Communist Party were to exert control over users' feeds through apps like TikTok.
- The potential TikTok ban or ByteDance divestiture is a response to concerns about national security and foreign influence, particularly from China. The bill aims to address the imbalance of access between U.S. and Chinese applications, with a focus on reciprocity in market access. There are fears that TikTok, under ByteDance's ownership, could pose a security threat if the Chinese Communist Party exerts control over users' data. The discussions around the ban involve considerations of free speech, political discourse, surveillance capabilities, and the need for evidence of national security risks.
- TikTok, a popular social media app owned by ByteDance, has faced scrutiny due to concerns that the Chinese government could influence or control the platform's content. This raises worries about potential data privacy breaches and the manipulation of information that could impact national security. The fear is that TikTok could be used to spread propaganda or conduct surveillance on users, leading to calls for increased scrutiny and potential restrictions on the app's operations in the United States. The debate revolves around balancing the app's popularity with the need to address these security and foreign influence risks.
- The discussions about reciprocity between U.S. and Chinese applications revolve around the unequal access that American apps have in China compared to Chinese apps in the U.S. This imbalance raises concerns about fairness and control over information flow between the two countries. The debate highlights how China restricts access to many American platforms while Chinese apps like TikTok have significant reach in the U.S. Calls for mutual access agreements are made to address this disparity.
- The fears related to free speech, particularly regarding Israel/Palestine issues, stem from concerns that the proposed legislation targeting apps like TikTok could limit discussions on these sensitive topics. Some worry that the bill might suppress voices advocating for Palestinian rights or criticizing Israeli policies, potentially impacting political discourse on social media platforms. The ambiguity in the bill's language and its potential to empower government actions against content related to Israel/Palestine could lead to fears of censorship and restrictions on free expression. These concerns highlight the complex intersection of national security considerations and the protection of free speech in the context of geopolitical tensions.
- TikTok, a popular social media app owned by ByteDance, has faced scrutiny over potential security risks due to its Chinese ownership. Concerns revolve around the possibility of the Chinese Communist Party (CCP) using TikTok to access and control user data, potentially leading to privacy breaches and national security threats. The fear is that the CCP could influence or manipulate content on TikTok to serve its interests, impacting political discourse and public opinion. These concerns have prompted discussions about the need for stricter regulations and oversight to mitigate these security risks.
- The skepticism about TikTok's code reliability and hosting arrangements stems from concerns about potential security vulnerabilities and data privacy risks associated with the app's ownership by ByteDance, a Chinese company. There are worries that the Chinese government could influence or access user data through the app. Moving TikTok's hosting to a U.S.-based company like Oracle is seen as a potential solution to mitigate these concerns. The discu ...
Counterarguments
- The bill could be seen as an overreach of government power, potentially setting a precedent for the government to ban or control private businesses based on national security without transparent criteria.
- Banning TikTok might not effectively address the broader issue of data privacy and security, as users may simply move to other platforms that could have similar vulnerabilities.
- The focus on TikTok could distract from the need for comprehensive data privacy legislation that applies to all companies, both foreign and domestic.
- The argument for reciprocity might not take into account the global nature of the internet, where restricting access based on the policies of another country could lead to a fragmented and censored web.
- The concerns about stifling free speech could be valid, as the ban might limit the diversity of platforms available for political discourse and expression, particularly for younger demographics who predominantly use TikTok.
- The potential security threat posed by TikTok could be mitigated through stringent data protection measures and oversight rather than an outright ban.
- Skepticism about TikTok's willingness to move hosting to the U.S. may not consider the company's efforts to comply with U.S. regulations and its agreement to store American user data on Oracl ...
Related Shortform Content
Get access to the context and additional materials
Opening AI Training Data and Derivative Works
Exploration of OpenAI's Training Practices
Potentially Unlicensed Use of YouTube Content
ChatGPT may have been trained with content from YouTube without licenses. This practice raises concerns about intellectual property (IP) rights, as it is unclear whether the necessary permissions were obtained for using these videos. Although the OpenAI CTO refrained from detailing the data sources for training ChatGPT, the model's response when the microphone is activated without speech—"thank you for watching"—suggests YouTube videos were part of the training dataset. OpenAI's use of such data might be construed as creating derivative works, further complicating the IP landscape.
Copyright and Legal Challenges
OpenAI faces a lawsuit from the New York Times over the issue of derivative works and licensing. David Friedberg asserts that training models with data from open internet sources like YouTube shouldn't be considered as generating derivative works. Conversely, David Sacks maintains that OpenAI possibly used publicly available data und ...
Here’s what you’ll find in our full summary
Opening AI Training Data and Derivative Works
Additional Materials
Clarifications
- Derivative works in the context of AI training involve creating new content based on existing copyrighted material. When AI models are trained using data like YouTube videos, the resulting model output can be seen as a derivative work. This raises questions about intellectual property rights and whether proper permissions were obtained for using the original content. Legal disputes can arise when the use of such data is contested by original content creators or other stakeholders.
- The fair use doctrine allows limited use of copyrighted material without permission for purposes such as criticism, commentary, news reporting, teaching, scholarship, or research. In the context of AI training, fair use could apply if the use of copyrighted material is transformative, meaning it adds new expression, meaning, or message to the original work. Fair use is a legal defense that balances the rights of copyright holders with the public interest in accessing and using creative works. Determining fair use involves considering factors like the purpose of the use, the nature of the copyrighted work, the amount used, and the effect on the market for the original work.
- The legal challenges surrounding AI training data involve concerns about intellectual property rights when using content from sources like YouTube without clear licenses. Disputes arise over whether training AI models with publicly available data constitutes creating derivative works and if fair use doctrines apply. Lawsuits, like the one between OpenAI and the New York Times, highlight the complexities of determining ownership and rights in the context of AI training practices. Questions also arise about the expertise of legal entities in handling technical nuances related to copyright issues in AI development.
- The technical expertise of legal actors in complex c ...
Counterarguments
- The use of data from platforms like YouTube for AI training could be considered transformative, which is one of the factors in determining fair use, potentially legitimizing OpenAI's practices.
- The concept of derivative works in the context of AI training data is still legally ambiguous, and until there is clear legal precedent, it may be premature to conclude that OpenAI's practices are infringing.
- The lawsuit from the New York Times may lead to a legal clarification that could benefit the entire AI industry by setting boundaries for what constitutes fair use of data in AI training.
- The argument that training AI with data from the internet creates derivative works could stifle innovation and the advancement of technology if overly restrictive interpretations are adopted.
- The ownership of rights to derivative works is a complex issue that may require new legal frameworks to balance the interests of original IP owners and third-party developers.
- Legal actors, such as judges and juries, may not currently ...
Related Shortform Content
Get access to the context and additional materials
Emergence of Vertical AI Startups
Vertical AI startups are gaining attention for their targeted approach to developing AI applications, aimed at enhancing or even replacing specific professional roles.
Developing AI agents for specific skills, jobs, professional roles (doctors, lawyers, engineers, customer support, etc.)
These emerging companies are creating AI tools dedicated to specific fields such as legal, medical, tax, and customer support. Notable examples include Harvey for legal counsel, Abridge for doctor’s note-taking, TaxGPT for tax assistance, and Sierra for customer support. These tools are designed to cater to the unique needs and terminologies of their respective professions.
Cognition's Devon shows impressive ability to fix bugs, tune models, develop apps. Rivals human engineers?
David Sacks and David Friedberg engage in a discussion about the application of Large Language Models (LLMs) like Cognition’s Devon in coding, which could radically increase productivity. Devon demonstrates an impressive set of skills, fixing bugs, tuning AI models, and building applications in real-time. Its coding competencies are showcased as superior to those of other generic language models, with the speculation that Devon might be built on top of GPT-4, with enhancements in reasoning and planning.
The technology, with its textual nature, aligns well with LLMs, allowing for potential high accuracy in tasks like debugging. Companies like Sourcegraph with product Kodi and GitHub’s Copilot seek to integrate AI with existing codebases.
Still several years from full autonomy and ability to replace human roles, but clear potential
Despite these advancements, full autonomy in AI replacing human roles might still be years away. However, the potential is palpable. Friedberg envisions AI agents expanding human potential, such as lawyers having dozens of AI associates or software engineers with many AI agents executing tasks. This enhancement ...
Here’s what you’ll find in our full summary
Emergence of Vertical AI Startups
Additional Materials
Clarifications
- Large Language Models (LLMs) are advanced AI systems capable of processing and generating human-like text. Cognition's Devon is a specific LLM known for its proficiency in tasks like fixing bugs, tuning AI models, and developing applications. These models leverage vast amounts of text data to understand and generate language, enabling them to perform a wide range of natural language processing tasks with high accuracy. Devon's capabilities showcase the potential of LLMs in enhancing productivity and efficiency in various fields by automating complex language-based tasks.
- Generative Pre-trained Transformer 4 (GPT-4) is a large multimodal language model developed by OpenAI, following the success of its predecessors like GPT-3. It was designed to predict the next token in a sequence by leveraging pre-training on vast amounts of data. GPT-4 introduced improvements over GPT-3.5, offering enhanced capabilities and larger context windows for processing text and images. OpenAI has not disclosed specific technical details about GPT-4, such as its exact model size.
- GitHub's Copilot is an AI-powered code completion tool developed by GitHub and OpenAI. It assists developers by generating code suggestions as they write, helping to speed up the coding process. Copilot is designed to work with various integrated development environments (IDEs) and supports multiple programming languages like Python, JavaScript, TypeScript, Ruby, and Go. It was first announced in June 2021 and has since evolved to incorporate advanced features like a chatbot based on GPT-4 an ...
Counterarguments
- Vertical AI startups may face significant ethical and regulatory challenges, especially in sensitive fields like legal and medical, where the stakes of decisions are high.
- AI tools may not be able to fully capture the nuances and complexities of human professions, leading to potential errors or oversights.
- The reliance on AI for professional roles could exacerbate issues of unemployment or underemployment among skilled workers.
- The effectiveness of AI agents like Devon in coding tasks may be overstated, as human oversight is often required to ensure quality and address edge cases.
- The assumption that Devon's coding competencies rival those of human engineers may not account for the creative and complex problem-solving abilities of experienced developers.
- The integration of AI with existing codebases could lead to new types of vulnerabilities and security concerns that need to be addressed.
- The prediction of full autonomy in AI replacing human roles may be overly optimistic, underestimating the adaptability and evolution of human job functions.
- The potential for AI to expand human potential in various industries may not be realized if there is a lack of investment in human-AI collaboration and training.
- The vision of companies run by single individuals with AI may overlook the importance of human collaboration, ...
Related Shortform Content
Get access to the context and additional materials
Florida Banning Lab-Grown Meat
Florida's state legislature is considering taking an unprecedented step by enacting a ban on lab-grown meat products, which has sparked a significant debate over the influence of incumbent ranchers and the implications for consumer choice and technological innovation.
Incumbent ranchers lobbying and regulatory capture
The panel discusses a bill passed by Florida's state legislature, which awaits the governor's signature, aiming to prohibit the manufacturing, sale, holding, or distribution of cultivated meat. Offenders could be charged with a second-degree misdemeanor. The bill is widely interpreted as a form of protectionism, with incumbent meat producers, notably ranchers, believed to be the driving force behind the legislative action. This situation suggests a classic case of regulatory capture, in which industry incumbents lobby for regulations that serve their interests at the expense of competition and innovation.
Restricting consumer choice and technological progress
The potential ban on lab-grown meat is seen as a significant restraint on consumer choice and a barrier to the progress of new and innovative technology within the food industry. David Friedberg likens the situation to imaginary past scenarios where other disruptive technologies could have been suppressed to protect incumbent industries, thereby highlighting how such regulatory actions hinder innovation.
...
Here’s what you’ll find in our full summary
Florida Banning Lab-Grown Meat
Additional Materials
Clarifications
- Regulatory capture occurs when regulators are influenced by specific industries or groups, prioritizing their interests over the public good. This can lead to biased regulations that benefit a small group at the expense of broader societal interests. The theory suggests that powerful stakeholders can manipulate regulatory decisions to serve their own agendas, undermining the intended purpose of regulation. Regulatory capture is a risk inherent to regulatory agencies, highlighting the importance of safeguarding regulators from undue influence.
- Lab-grown meat, also known as cultured or cultivated meat, is produced by growing animal cells in a lab setting rather than raising and slaughtering animals. This technology aims to provide an alternative to traditional animal agriculture, offering a more sustainable and ethical way to produce meat. The process involves taking a small sample of animal cells, such as muscle cells, and nurturing them in a controlled environment to grow into meat that is biologically identical to conventionally produced meat. Lab-grown meat has the potential to address environmental concerns, animal welfare issues, and food security challenges associated with traditional meat production.
- Protectionism in the context of the food industry involves policies or actions that shield domestic producers from foreign competition. In this case, the potential ban on lab-grown meat can be seen as a form of protectionism to protect traditional meat producers like ranchers from competition. It aims to maintain the market dominance of incumbent players by restricting the entry of new technologies or products that could challenge their position. P ...
Counterarguments
- The ban may be intended to protect consumers from untested and potentially unsafe food products, as the long-term health effects of consuming lab-grown meat are not yet fully understood.
- The legislation could be seen as a way to preserve traditional agricultural practices and protect the livelihoods of ranchers who may not be able to compete with lab-grown meat producers on price or scale.
- Some may argue that the state has a right to exercise caution and regulate new technologies to ensure they align with the values and preferences of its citizens.
- There could be concerns about the environmental impact of lab-grown meat production, such as energy consumption and waste management, which might justify regulatory scrutiny.
- The ban might be a temporary measure to allow for a more comprehensive regulatory framework to be developed that ensures lab-grown meat is safe, ethically produced, and properly labeled.
- It could be argued that the market for lab-grown meat is not yet mature enough to require the same level of consumer choice as other food products, and that more research and ...
Related Shortform Content
Get access to the context and additional materials
Create Summaries for anything on the web
Download the Shortform Chrome extension for your browser
