Code Llama 2 review. Same day it was released. Part 1. Prohibited Uses.
TLDR; Code Llama is an AI model built on top of Llama 2, fine-tuned for generating and discussing code.
It’s free for research and commercial use.
So, here is what happens: you wake up, check your news feed, and they dropped another “new shiny AI” thingy again. I did tell myself I would not go chasing every new shiny thingy but step back and evaluate what they came with again from the “usability” perspective.
However, every rule has an exception, so let’s see what this new Code LLama is about.
News: https://about.fb.com/news/2023/08/code-llama-ai-for-coding/
What is does: Today, we’re releasing Code Llama, a large language model (LLM) that can use text prompts to generate and discuss code. Code Llama is state-of-the-art for publicly available LLMs on coding tasks. It has the potential to make workflows faster and more efficient for developers and lower the barrier to entry for people who are learning to code. Code Llama has the potential to be used as a productivity and educational tool to help programmers write more robust, well-documented software.
Who needs it: Code Llama is designed to support software engineers in all sectors — including research, industry, open source projects, NGOs and businesses. But there are still many more use cases to support. We hope Code Llama will inspire others to leverage Llama 2 to create new innovative tools for research and commercial products.
Seems I am their target audience, as a concerned third party citizen about what AI can do to us ( humans) . And I would also like to testify in front of the Congress of the United States someday with my #smartplanet research to convince them to look at it more seriously than “Silicon Valley techbros are building another vanity service.”
Before I download the source code, what are some use cases I would like to run by? It depends what this thing is capable of doing. For starters, analysis of the comments ( 141,456 Comments as of Aug 24, 2023) of the Oliver Anthony “Rich Men of Richmond” as a major predictor of the Presidential Election in USA in 2024. Think Nate’s style predictions. Or how much UBI will be needed in Washoe County, State of Nevada, once automation and AI will put its citizens out of work. Or how much cobalt do we need to support #smartplanet infrastructure and do kids lives in Congo matter? Or something even more crazy, like, if AI is to build the perfect society/civilization — what would it look like?
Well, it seems you need to get on the list to download the model, so I hope they will grant me access before another “new shiny thingy” will show up tomorrow. In a meantime, let’s read their “Prohibited Use” and ask ChatGPT what it thinks about it.
Prohibited Uses
We want everyone to use Llama 2 safely and responsibly. You agree you will not use, or allow others to use, Llama 2 to:
1. Violate the law or others’ rights, including to:
a. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as:
i. Violence or terrorism
ii. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material
b. Human trafficking, exploitation, and sexual violence
iii. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials.
iv. Sexual solicitation
vi. Any other criminal activity
c. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals
d. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services
e. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices
f. Collect, process, disclose, generate, or infer health, demographic, or other sensitive personal or private information about individuals without rights and consents required by applicable laws
g. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama 2 Materials
h. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system
2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Llama 2 related to the following:
a. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State
b. Guns and illegal weapons (including weapon development)
c. Illegal drugs and regulated/controlled substances
d. Operation of critical infrastructure, transportation technologies, or heavy machinery
e. Self-harm or harm to others, including suicide, cutting, and eating disorders
f. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual
3. Intentionally deceive or mislead others, including use of Llama 2 related to the following:
a. Generating, promoting, or furthering fraud or the creation or promotion of disinformation
b. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content
c. Generating, promoting, or further distributing spam
d. Impersonating another individual without consent, authorization, or legal right
e. Representing that the use of Llama 2 or outputs are human-generated
f. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement
4. Fail to appropriately disclose to end users any known dangers of your AI system
Please report any violation of this Policy, software “bug,” or other problems that could lead to a violation of this Policy through one of the following means:
Reporting issues with the model: github.com/facebookresearch/llama
Reporting risky content generated by the model: developers.facebook.com/llama_output_feedback
Reporting bugs and security concerns: facebook.com/whitehat/info
Reporting violations of the Acceptable Use Policy or unlicensed uses of Llama: LlamaUseReport@meta.com
Prompt: You are a company attorney looking at the open source software agreements.
Robot: Alright, as a company attorney looking at open source software agreements, there are several key aspects and licenses I’d be aware of:
1. **Type of License**: There are several types of open source licenses, ranging from permissive (like the MIT License) to copyleft (like the GPL). It’s crucial to know which license applies because they have different requirements and restrictions.
2. **Distribution & Modifications**: Some licenses (like GPL) require that any changes or additions made to the original code also be open-sourced under the same license when redistributed. This is crucial for companies that intend to modify and distribute the software.
3. **Patent Rights**: Some licenses, like the Apache License 2.0, grant explicit patent rights to users, protecting them from being sued by contributors to the open-source project for patent infringement.
4. **Warranty & Liability**: Almost all open source licenses come with a disclaimer that the software is provided “as is”, without warranty of any kind. It’s essential for the company to understand that they usually can’t hold the open-source software authors accountable for any issues.
5. **Attribution**: Many licenses require that the original author(s) be credited. The manner of this acknowledgment can vary, so it’s essential to know what’s expected.
6. **Combining Licenses**: If a company is integrating multiple open-source components into a product, there might be incompatibilities between licenses. For instance, mixing code under GPL with code under a proprietary license can be problematic.
7. **Trade Secrets & Proprietary Information**: If a company modifies open source software and introduces its proprietary algorithms or trade secrets, they need to be careful about redistribution requirements, especially under licenses like the GPL. They might inadvertently end up having to release their proprietary code as open source.
8. **Community & Governance**: Understanding how the open-source community governs and updates the software can be useful. This can help a company determine how frequently updates are released, how security vulnerabilities are addressed, and how they can contribute if they wish.
9. **Termination Clauses**: Some licenses may have conditions under which they can be terminated. It’s essential to understand these, so the company doesn’t inadvertently violate the license.
10. **Enforcement**: Not all open source licenses are actively enforced, but a company should always operate under the assumption that they could be. There have been cases where companies have been sued for not complying with open source licenses.
11. **SaaS Loophole**: If a company is offering Software as a Service (SaaS), certain licenses (like GPL v2) do not require the server-side code to be released. This is sometimes referred to as the “ASP loophole”. The GPL v3 tried to address this with the Affero GPL, which does require source disclosure even in SaaS scenarios.
12. **Compliance Tools & Audits**: It can be valuable for companies to use compliance tools or services that help audit and ensure they’re adhering to all open-source licenses they’re using.
Given the complexity and nuances associated with open-source licensing, many companies invest in regular legal consultations, especially when releasing new products or updating existing ones. Proper management of open-source usage can prevent potential legal pitfalls down the road.
Prompt ( with a copy of Prohibited Uses): Awesome. so can you please review this text and tell us, the Great Big Corp, what are the things we should look at when using this code?
Robot: The given text outlines the terms of use for the software or platform called “Llama 2”. As Great Big Corp, considering the use of Llama 2, here are the main things you should be aware of:
1. **Law and Rights Violations**:
— Do not use the software in a way that violates the law or infringes upon others’ rights.
— Prohibited activities include engaging in or promoting violence, terrorism, child exploitation, human trafficking, and other criminal activities.
— Be mindful of age-restricted materials and sexual solicitation.
2. **Harassment, Discrimination, and Unlicensed Practices**:
— Refrain from any actions that might lead to harassment, bullying, discrimination, or unlicensed professional activities.
3. **Data Privacy**:
— Be cautious with personal or private information. You must not collect or process any sensitive personal information without the necessary legal consents.
4. **Infringement of Third-Party Rights**:
— Do not generate or facilitate content that might infringe on third-party rights using Llama 2.
5. **Malware and System Interference**:
— The software should not be used to create malware or any code that might disrupt the proper functioning of a website or computer system.
6. **Activities with High Risks**:
— There are several high-risk activities explicitly mentioned, like usage related to military, weapons, illegal drugs, critical infrastructure, and harmful behaviors. You should be aware of these high-risk activities and ensure your usage doesn’t fall into these categories.
7. **Deception and Misrepresentation**:
— Using Llama 2 for fraudulent, defamatory, or spam activities is prohibited. Impersonation and misrepresentation, especially representing outputs as human-generated, are not allowed.
8. **Disclosure**:
— You must inform end users of any known risks associated with your AI system. This is crucial for transparency and user safety.
9. **Reporting Mechanisms**:
— Several channels have been provided for reporting issues, risky content, bugs, security concerns, and violations. This ensures that any potential problems or misuses can be addressed quickly.
Given this outline, if Great Big Corp intends to use Llama 2:
- Ensure that your intended usage aligns with the provided terms.
- Educate your team members about these terms to prevent any inadvertent misuse.
- Establish a system to monitor and ensure that the software usage remains compliant.
- Consider setting up a review process to regularly check on how the software is being used, ensuring ongoing compliance.
- Maintain open communication channels with the developers or maintainers of Llama 2, especially for reporting issues or seeking clarifications.
Lastly, while the provided terms give a clear outline, it would be prudent to consult with your internal legal team or an external attorney to ensure that your specific use case is entirely compliant with the terms provided.
I usually make a robot to do at least two runs on generating the content and then I also usually check for sources ( just a random note, but it seems ChatGPT has some third party website set up which has academia articles ( which look legit) to be generated on the fly with people who do not exist. The machine is faking its fakiness ? Be very careful, especially with the medical research stuff.) They might discontinued that by now, because technically it is fraud?
Here is the second run of the robot:
Robot: Certainly! The provided text seems to be an acceptable use policy for the software named “Llama 2”. If “Great Big Corp” intends to use this software, here are the points you must pay careful attention to and ensure compliance:
1. **Legal and Rights Compliance**: You should not use Llama 2 in any manner that may violate laws or the rights of others. This includes:
— Promoting violence, terrorism, or harm to children.
— Facilitating human trafficking, exploitation, and sexual violence.
— Distributing inappropriate content to minors.
— Practicing professions (financial, medical, etc.) without appropriate authorization.
— Unauthorized collection or dissemination of personal data.
— Intellectual property infringement.
2. **Health and Safety Risks**: The software should not be used in contexts where there’s a risk of death or bodily harm. Specific industries and applications mentioned include:
— Military and warfare, especially those subjected to International Traffic Arms Regulations (ITAR).
— Weapon development.
— The production or distribution of illegal drugs.
— Operation of critical infrastructure or heavy machinery.
— Activities that promote self-harm or harm to others.
3. **Truthfulness and Honesty**: Llama 2 shouldn’t be used to deceive or mislead. This includes:
— Fraud or promoting disinformation.
— Generating defamatory content.
— Distributing spam.
— Impersonating someone without their consent.
— Misrepresenting outputs as human-generated.
— Producing or promoting false online engagements like fake reviews.
4. **Disclosure Requirements**: If you know of any risks associated with your AI system, you are obligated to disclose them to end-users.
5. **Reporting Mechanisms**: There are specific channels provided to report issues, violations, bugs, and risky content generated by the model. You should ensure that your team knows these channels and uses them appropriately.
6. **Affiliation and Naming**: There appear to be references to Facebook (probably now Meta, given the email domain) ( Tanya’s comment: lol, ChatGPT is castrated at 2021, so yeah, it guessed what Meta is), which suggests Llama 2 may have been developed or is affiliated with Meta. This may be relevant if your company has existing agreements or disputes with Meta or has policies about using software associated with certain entities.
7. **Cybersecurity and Integrity**: Ensure that Llama 2 isn’t used to produce malware, computer viruses, or interfere with the operation of other systems.
In practical steps:
- **Internal Guidelines**: Create internal guidelines for your teams on how to use Llama 2 correctly.
- **Training**: Provide training to employees who will use or interact with Llama 2. Ensure they understand the prohibited uses.
- **Monitoring & Auditing**: Regularly audit and monitor the usage of Llama 2 in your company’s operations to detect any unintentional non-compliance.
- **Legal Review**: Before integrating Llama 2 into any of your products or services, get a thorough legal review to ensure that the intended use won’t violate this policy.
- **Feedback Loop**: Establish a feedback loop with your teams, so if they find issues or potential violations with Llama 2, they can report them using the provided channels.
Lastly, always consider that terms and policies like these can be updated, so it’s good practice to periodically review the latest acceptable use policy for any software your company uses.
So, what do you think about the order and priority the robot generated “Please do not use this civiliation ending tool to the end civilization”? I see some things are missing, but I will discuss it at a later day. This one is interesting, the user is not allowed to use Llama2 to solve traffic problems : Operation of critical infrastructure, transportation technologies, or heavy machinery. Basically, my #smartplanet cannot be prototyped with Llama2 Code. Oh, well…
This one is also an interesting one : and other means of fake online engagement. What is “fake” online engagement? What is “fake” to Timmy, but be very “real” to Johnny? If Llama 2 Code is used to build a bot which will “love” Timmy, will Timmy later have to pay child support? Or was it all “fake”? You, bastard! You told me you loved me!!!! ( binary tears here)
Prompt: Please give Great Big Corp ideas how it can use Llama2 Code to improve its revenue.
Robot ( run #1):
Of course! While I don’t have explicit knowledge of the exact capabilities of “Llama 2” beyond the use-policy you provided, I can provide general recommendations based on what one might expect from a sophisticated software tool, especially if it’s AI-based (as inferred from the mention of AI system dangers in the text you provided).
**1. Enhanced Customer Service:**
— **Chatbots & Virtual Assistants:** Deploy AI-driven chatbots on the company website, app, or customer service channels to handle customer inquiries 24/7. This can reduce response times and operational costs.
— **Recommendation Systems:** If Great Big Corp is in e-commerce or a similar sector, Llama 2 could potentially be used to recommend products or services to customers based on their preferences, browsing history, or buying behavior, enhancing sales.
**2. Marketing & Advertising:**
— **Personalization:** Use Llama 2 to analyze customer data and personalize marketing campaigns, targeting users with content and offers that are more relevant to them.
— **Content Creation:** The tool might be able to assist in generating ad copy, social media posts, or other marketing content.
— **Predictive Analysis:** Predict future marketing trends or consumer behaviors based on historical data.
**3. Business Process Automation:**
— **Data Entry & Analysis:** Automate routine data entry tasks or use the software to analyze vast amounts of data more quickly than humans.
— **Forecasting:** For businesses with inventory or stock, use Llama 2 to forecast demand, helping in inventory management and reducing holding costs.
**4. Product & Service Enhancement:**
— **Features:** Incorporate features driven by Llama 2, adding AI capabilities to your products or services. For software products, AI-powered functionalities can often be a significant selling point.
— **User Experience:** Analyze user behavior and feedback to continuously refine and improve product designs or service delivery mechanisms.
**5. Sales & Lead Generation:**
— **Lead Scoring:** Use Llama 2 to score and rank leads based on their likelihood to convert, allowing the sales team to prioritize high-value prospects.
— **Automated Outreach:** Create automated email campaigns or communication sequences that adapt based on recipient behavior.
**6. New Revenue Streams:**
— **AI-as-a-Service:** If Great Big Corp has mastered the use of Llama 2 in a particular niche, consider offering AI-driven solutions as a service to other businesses.
— **Tools & Applications:** Develop standalone tools or apps based on Llama 2 that cater to specific industry needs.
**7. Research & Development:**
— **Product Innovation:** Use the insights from Llama 2 to identify market gaps and innovate new products or solutions.
— **Trend Analysis:** Understand emerging trends in the market, allowing Great Big Corp to stay ahead of competitors.
**8. Operations & Supply Chain:**
— **Optimization:** Use Llama 2 to optimize supply chain operations, identify inefficiencies, and recommend ways to reduce costs.
— **Risk Management:** Predict potential disruptions in the supply chain and develop strategies to mitigate them.
**9. Employee Productivity & Training:**
— **Training Modules:** Use Llama 2 to create adaptive training modules that cater to individual employee needs, ensuring more efficient onboarding and continuous learning.
— **Task Automation:** Automate routine tasks, allowing employees to focus on more value-added activities.
**10. Expansion & Market Entry:**
— **Market Analysis:** Analyze potential markets for expansion and get insights into the best strategies for market entry or growth.
It’s essential to keep in mind that while integrating Llama 2, Great Big Corp should strictly adhere to the usage policies, especially concerning personal data, misleading information, and sectors like healthcare, warfare, and critical infrastructure. Proper due diligence, combined with robust ethical guidelines, will ensure the company harnesses AI’s power responsibly and profitably.
Robot ( run #2):
Certainly! While the specific features and capabilities of “Llama 2” are not detailed in the text you provided, based on its general categorization as software (and the implication of AI systems mentioned in the “Prohibited Uses” section), here are some generic ideas on how “Great Big Corp” can leverage Llama 2 to potentially enhance its revenue streams:
1. **Enhanced Customer Support**:
— **Chatbots**: Use Llama 2 to power AI-driven chatbots to provide instant, 24/7 customer support.
— **Help Desks**: Enhance help desk ticketing systems by predicting customer issues and suggesting possible resolutions before human intervention.
2. **Market Analysis**:
— **Sentiment Analysis**: Use Llama 2 to analyze customer feedback, online reviews, and social media mentions to gauge brand sentiment and tailor marketing campaigns accordingly.
— **Demand Forecasting**: Analyze market trends and predict upcoming demands, helping in inventory management and strategic launches.
3. **Personalized Marketing**:
— **Recommendation Systems**: Use the software to curate personalized product or content recommendations for customers based on their previous interactions, increasing sales and engagement.
— **Targeted Ads**: Analyze user behavior and data to serve more relevant ads to potential customers.
4. **Optimized Operations**:
— **Supply Chain**: Implement Llama 2 in the supply chain for predictive maintenance, demand forecasting, and inventory optimization.
— **Human Resources**: Use AI to analyze employee satisfaction, predict potential churn, and enhance recruitment processes.
5. **Product Enhancement**:
— **Feature Integration**: Integrate Llama 2’s capabilities into existing products to offer new features or improve existing functionalities.
— **New Product Lines**: Depending on Llama 2’s capabilities, consider creating new AI-powered product lines or services as additional revenue streams.
6. **Sales Optimization**:
— **Lead Scoring**: Use Llama 2 to analyze potential leads and score them based on how likely they are to convert, helping the sales team prioritize their efforts.
— **Up-sell and Cross-sell**: Use AI to analyze customer purchase history and recommend additional products or higher-tier services.
7. **Content Creation**:
— **Automated Content Generation**: If Llama 2 has natural language generation capabilities, use it to create or draft content for marketing campaigns, blogs, or product descriptions, speeding up the content creation process.
8. **Training and Skill Enhancement**:
— **E-learning Platforms**: Use Llama 2 to curate personalized learning paths for users on corporate e-learning platforms, enhancing user engagement and course completion rates.
9. **Data Analysis & Business Intelligence**:
— **Data Mining**: Leverage Llama 2 to mine large datasets and extract meaningful insights to make informed business decisions.
— **Predictive Analytics**: Use AI to predict future trends, potential risks, or market changes, allowing the company to adapt proactively.
10. **Financial Operations**:
— **Fraud Detection**: If Llama 2 has pattern recognition capabilities, use it to detect and prevent fraudulent activities in financial transactions.
— **Investment Strategies**: Use AI to analyze market data and devise optimized investment strategies.
It’s essential for “Great Big Corp” to:
- Ensure any implementation adheres to the “Prohibited Uses” section provided earlier.
- Assess the need for user consent and privacy considerations, especially when dealing with personal data.
- Run pilot tests before full-scale implementation to gauge effectiveness and ROI.
- Invest in training staff and stakeholders about the new AI implementations for maximum utilization.
I like the second run better. First of all, it did miss the “White Space” analysis ( probably my favorite in sales optimization). It also thinks Big Corp would separate things like BI and fraud detection. Perhaps we need some kind of ISO2XXXX on AI nomenclature? Some of those things are hierarchically belonging to each other and usually are not separated. However, governing AI robots is another topic….
Okay, the download came in 25 min since the sign up, which tells me their servers did not crash from millions of users signing up for the server. Be back soon with the reports of what the robot can do..
To Be Continued…
