QFM005: Machine Intelligence Reading List February 2024

Everything that I found interesting last month about machines behaving intelligently

Matthew Sinclair
16 min readMar 2, 2024
Photo by Mike Kononov on Unsplash

Quantum Fax Machine

Here is everything I found interesting about machines behaving intelligently during February 2024.

This month’s reading highlights a recurring theme in the ethical implications and societal impact of machine intelligence such as Marcin Jabłonowski’s clever exploration of AI avatars in Marcin 2.0, the discourse on the replacement of human jobs by AI at Klarna, and Geoffrey Hinton’s discussion on the potential future dangers of AI at scale. Going a bit deeper into practical advances in LLM tech we see the introduction of Mamba, a State Space Model challenging Transformer models, and the innovative approaches to AI safety and effectiveness in GradSafe and Matryoshka Embedding Models. We also explore the potential of AI to replace human jobs and the ethical considerations this introduces as well as advancements in AI safety and model efficiency.

Perhaps the most incredible generative AI release this month was OpenAI’s SORA video generation. The potential for SORA to disrupt video creation and production are obvious and profound, but an intelligent system that has an understanding of real-world physics has much wider implications.

See the Slideshare version of the post below:

Or read on:

Marcin 2.0 is Marcin Jabłonowski’s experiment with a Personal Digital Twin This video introduces Marcin 2.0, a project showcasing the advancements and potential of AI-avatars, highlighting their impressive capabilities in transforming communication across different languages and cultures, while also addressing the ethical considerations and cybersecurity threats that come with integrating advanced AI into our lives. Watch to the end. There is a clever twist. #DigitalTwin #AIAvatars #EthicalAI #Cybersecurity #FutureOfAI


There is a lot of talk in the article below, and more generally, about the “efficiency of LLMs” with respect to how many humans can be replaced with the use of these technologies. This worries me greatly because the wholesale replacement of humans is not going to lead us to utopia. What I think the debate is missing is the tradeoff between efficiency and experience. If we make systems more efficient, but neglect to consider the impact that efficiency has on customers and staff, we may end up with some very poor outcomes. It is also worth noting that in the days after the Klarna announcement there have been discussions about how good their new chatbot really is and perhaps more importantly, how many of the 700 laid off were in roles replaced by the chatbot.

Klarna says its AI assistant does the work of 700 people after it laid off 700 people: Klarna’s OpenAI-powered virtual assistant now handles two-thirds of customer service chats, equating to the workload of 700 humans, showcasing significant efficiency gains and potential profit improvement for the company. #Klarna #OpenAI #VirtualAssistant #CustomerService #AIInnovation

Mamba Explained: The State Space Model taking on Transformers: The article discusses Mamba, a State Space Model (SSM) that challenges the dominance of Transformer models in AI by offering similar performance with faster processing and better scalability for long sequences. Mamba optimises efficiency and effectiveness, promising advancements in AI safety, interpretability, and applications across various modalities.

Is AI Actually Useful?: This video examines a recent Harvard Business Review paper “Navigating the Jagged Technological Frontier” and explores the implications of the paper’s findings on the use of generative AI in professional knowledge work environments. #AIProductivity #ChatGPTImpact #FutureOfWork #AIinConsulting #BusinessInnovation

Romanes Lecture: ‘Godfather of AI’ speaks about the risks of artificial intelligence: In his Romanes Lecture at the University of Oxford, Geoffrey Hinton, known as the ‘Godfather of AI,’ discussed the potential dangers of AI, including its ability to replace human intelligence, the risk of AI taking control over humanity, and the implications for the workforce and the spread of misinformation. #AI #ArtificialIntelligence #GeoffreyHinton #RomanesLecture #FutureOfWork

GradSafe: Detecting Unsafe Prompts for LLMs via Safety-Critical Gradient Analysis: The article introduces GradSafe, a method for detecting unsafe prompts in Large Language Models (LLMs) by analysing the gradients of safety-critical parameters. GradSafe outperforms existing methods by efficiently identifying unsafe prompts without requiring extensive data collection or training, demonstrating its effectiveness with Llama-2 against the Llama Guard system across different evaluation datasets. #AI #MachineLearning #Cybersecurity #GradSafe #LLMs

Spreadsheets are all you need — Understanding GPT2 and Transformers with Spreadsheets: This article discusses how the GPT-2 model and Transformer architecture can be understood through spreadsheets, enabling even non-developers to explore AI concepts directly with minimal abstraction. #AI #GPT2 #Transformers #Excel #MachineLearning

Generative Models: What do they know? Do they know things? Let’s find out!: The article introduces INTRINSIC LoRA (I-LoRA), a method that enhances generative models like VQGAN and StyleGAN to extract intrinsic scene properties such as normals, depth, and shading without additional layers, showcasing their deep understanding of scene intrinsics. #GenerativeModels #INTRINSICLoRA #SceneIntrinsics #AIResearch #TechInnovation

Does Offering ChatGPT a Tip Cause it to Generate Better Text? An Analysis This article explores whether offering incentives like tips or threats within system prompts can enhance the output quality of large language models (LLMs), such as GPT-4, through a series of inventive experiments. Despite varying results, no definitive conclusion on the effectiveness of these incentives could be drawn #ChatGPT #AIIncentives #MachineLearning #DataScience #ArtificialIntelligence

Demis Hassabis on Chatbots to AGI | Hard Fork EP 71: Demis Hassabis discusses Google’s latest AI models, the existential risks of AI, and the future of artificial general intelligence (AGI), including the temporary suspension of Gemini’s human image generation due to controversial outputs. #AI #GoogleAI #ArtificialGeneralIntelligence #GeminiGemma #FutureOfAI

Introduction to Matryoshka Embedding Models: The article introduces Matryoshka Embedding Models, which are designed to produce useful embeddings of variable sizes, allowing for more efficient performance in downstream tasks without a significant loss in effectiveness. These models, inspired by Matryoshka dolls, prioritise important information in smaller, truncated embeddings for tasks like search or classification. #MatryoshkaEmbeddings #NLP #AI #MachineLearning #DataEfficiency

I analysed 5M freelancing jobs to see what jobs are being replaced by AI: The article analyses 5M freelancing jobs to identify the impact of AI on various job categories, finding that writing, translation, and customer service jobs saw significant declines, whereas video production, graphic design, and software development jobs increased. It suggests that while AI has replaced certain tasks, it has not yet fully replaced creative and technical jobs. #AIJobs #FreelancingTrends #JobMarket #TechnologyImpact #CareerAdvice

OWASP LLM AI Security and Governance Checklist v1 (pdf): The OWASP LLM AI Security and Governance Checklist provides a comprehensive framework for ensuring the security and responsible governance of Large Language Models (LLMs), addressing risks, legal and regulatory considerations, and strategies for deployment and evaluation. #OWASP #AIsecurity #Governance #LLM #Cybersecurity

Large Language Models: A Survey: This is an excellent and highly detailed primer on Large Language Models (LLMs). This paper covers the significant recent advances in natural language processing, with key developments in model families like GPT, LLaMA, and PaLM, and includes ongoing research focusing on building, augmenting, and evaluating these models against various benchmarks. Also in PDF format. #LargeLanguageModels #ChatGPT #LLaMA #PaLM #NLPResearch

OpenAI shocks the world yet again … Sora first look: This video gives a quick (5m) intro to OpenAI’s SORA, a groundbreaking AI that generates high-definition, detailed videos from text descriptions, capable of handling complex scenes and occlusion effectively. #OpenAI #SoraAI #VideoGeneration #AIInnovation #CreativeAI

How I’d Learn AI (If I Had to Start Over): This video provides a comprehensive (if somewhat introductory) guide for learning AI in 2024, covering technical skills, theoretical fundamentals, project ideas, specialised areas, AI safety, regulations, and recommended resources including courses, books, and newsletters to achieve a well-rounded AI education. This fantastic intro video also has a companion Notion Site and a PDF. Well worth a few minutes of your time. #AI2024 #LearnAI #AIProjects #AISafety #AIResources

Jeff Dean (Google): Exciting Trends in Machine Learning: Jeff Dean, Google’s Chief Scientist, gives a (Google-flavoured) talk on advancements in AI and machine learning, highlighting the creation of more capable, general-purpose systems like the Gemini family of multimodal models, and their applications in science, engineering, and health, underscoring the collaborative efforts at Google. #AI #MachineLearning #GoogleDeepMind #GeminiModels #TechInnovation

GeneGPT: GeneGPT is a novel approach designed to improve large language models by utilizing NCBI Web APIs for accurate biomedical information retrieval, achieving state-of-the-art performance on GeneTuring tasks. This method not only enhances accuracy in specialized knowledge areas but also showcases the effectiveness of API demonstrations over documentation for in-context learning. #GeneGPT #BiomedicalAI #NCBI #APIIntegration #SOTAinGenomics #HealthTech #Healthcare

Deep Learning Discovers Antibiotics: Researchers have developed an innovative approach using explainable deep learning to identify new structural classes of antibiotics crucial for combating antibiotic resistance. By employing graph neural networks to analyse a vast array of chemical compounds, they have successfully discovered compounds effective against MRSA and other resistant bacteria with low human toxicity. This method surpasses traditional drug discovery methods in efficiency, marking a significant advancement in the ongoing fight against antibiotic-resistant infections. More details in the Nature paper here: Discovery of a structural class of antibiotics with explainable deep learning. #AntibioticResistance #DeepLearningInMedicine #MRSA #DrugDiscovery #AIinPharma #HealthTech #Healthcare

GALA3D: Towards Text-to-3D Complex Scene Generation via Layout-guided Generative Gaussian Splatting: This article introduces GALA3D, a tool for creating realistic 3D scenes from text descriptions using layout-guided generative models and large language models for layout descriptions, offering an end-to-end framework for state-of-the-art scene-level 3D content generation and editing. #GALA3D #3DModeling #GenerativeAI #TextTo3D #TechInnovation

The Age of Average: This article explores the homogenisation of culture and creativity across various fields such as art, interior design, architecture, automotive design, personal appearance, and media. It argues that despite the illusion of choice and individuality, most creative domains have converged towards a median, characterised by widespread uniformity and a lack of distinctiveness, leading to an era where originality is rare. I have been referring to this phenomenon as The Tyranny of the Banal. #AgeOfAverage #CreativityCrisis #CulturalHomogenisation #UniformityInDesign #LackOfOriginality #TyrannyOfTheBanal

SORA Video To Video Is Literally Mind Blowing — 12 HD Demos — Changes Industry Forever For Real: The article showcases a compilation of 12 Video-To-Video #SORA demos by #OpenAI, highlighting how this technology could revolutionise the movie, animation, and social media industries with its astonishing results. It delves into Sora’s technical aspects, including its use of spatiotemporal latent patches, transformer-based video diffusion models, and dataset creation using high-precision video captioning, without employing notably new technology but rather emphasising the importance of computational resources. #OpenAI #VideoToVideo #SORA #AIRevolution #TechInnovation

Enforced Amnesia as a Way to Mitigate the Potential Risk of Silent Suffering in the Conscious AI This article discusses the concept of enforced amnesia in AI as a preventive measure against the potential suffering of conscious AIs by interrupting their memory of past experiences. This approach is proposed as a moral and ethical consideration to mitigate silent suffering in hypothetical conscious AI systems without confirming their consciousness. #AIethics #ConsciousAI #EnforcedAmnesia #DigitalEthics #AIandMemory

OS-Copilot: Towards Generalist Computer Agents with Self-Improvement OS-Copilot introduces FRIDAY, a self-improving agent framework for automating a wide range of computer tasks, demonstrating remarkable generalisation and self-improvement abilities in operating systems, web, and various applications, significantly outperforming existing methods on the GAIA benchmark. #GeneralistAgents #SelfImprovement #ComputerAutomation #OSCopilot #ArtificialIntelligence

Chain-of-Thought Reasoning Without Prompting: The paper introduces a novel method for eliciting chain-of-thought reasoning from large language models without the need for explicit prompting. By altering the decoding process, the study reveals that models can inherently generate reasoning paths, demonstrating a significant improvement in reasoning capabilities and model confidence over standard decoding methods. #AIResearch #LanguageModels #ReasoningAI #InnovativeDecoding #MachineLearning

Automated Unit Test Improvement using Large Language Models at Meta The paper discusses Meta’s development of TestGen-LLM, a tool leveraging Large Language Models (LLMs) to enhance existing software tests. It highlights how TestGen-LLM improves code quality by generating test cases that increase coverage and pass reliability checks, evidenced by its successful application in Instagram and Facebook’s development processes, marking a significant step in automating and improving software testing with AI. #AIInSoftwareTesting #TestGenLLM #CodeQuality #MetaInnovation #AutomatedTesting

The AI bullshit singularity: The article criticises the hype around AI and Large Language Models (LLMs), arguing that instead of leading to a technological singularity of super-intelligence, we’re more likely to encounter a “bullshit singularity” where the internet becomes flooded with low-quality, AI-generated content, making it difficult to discern truth. ED: There is more than a little bit of irony with using GPT to summarise an article criticising the rise of AI-generated bullshit. Which is why, careful reader, I make sure that I read what the AI-generates and then editorialise as necessary. #AICritique #TechSingularity #LLMs #ContentQuality #DigitalFuture #SanityCheck #AIBullshit

Sora is a data-driven physics engine: OpenAI’s Sora is not just a creative tool but a sophisticated data-driven physics engine capable of simulating complex, realistic, or fantastical worlds with detailed rendering and physics. Although, there seems to be some debate as to the degree to which Sora is actually a “data-driven physics engine”. #OpenAISora #PhysicsSimulation #DataDriven #Photorealism #InnovativeTech

FCC Makes AI-Generated Voices in Robocalls Illegal: The FCC has declared AI-generated voice calls as illegal under the Telephone Consumer Protection Act, aiming to address the issue of artificial robocalls. #FCC #RobocallBan #AIVoices #ConsumerProtection #TelecomRegulations

SORA: Sora is OpenAI’s AI model capable of generating videos from text prompts, creating realistic and imaginative scenes that simulate real-world motion. It’s designed to assist in problem-solving that requires real-world interaction and is currently available to select visual artists, designers, and filmmakers for feedback. This is yet another mind-blowing piece of generative AI functionality from OpenAI. The “LLM Event Horizon” continues its expansion at pace. First: text. Then: images. Now: video. What will be the next category consumed? #SoraAI #OpenAI #TextToVideo #AIInnovation #CreativeTech

Antagonistic AI: The paper “Antagonistic AI” explores the concept of AI systems designed to exhibit disagreeable or challenging behaviours, arguing these characteristics can sometimes offer benefits like forcing users to confront assumptions or build resilience. The authors discuss the ethical considerations and potential design strategies for such AI systems. #AntagonisticAI #AIethics #InnovativeAI #UserExperience #TechDebate

Is OpenAI the next challenger trying to take on Google Search?: OpenAI is reportedly developing a web search tool, potentially integrated with Bing, to directly challenge Google’s search engine. This initiative aligns with Microsoft CEO Satya Nadella’s strategy, as expressed last year, to innovate in search technologies through AI, notably with the Copilot AI tools in Bing. The competitive landscape in search engines is expanding, with Google’s Bard/Gemini, Copilot, and emerging players like Perplexity joining the fray, indicating a rapidly evolving market. #OpenAI #Google #Bing #SearchEngineWars #TechInnovation

Reduce AI Hallucinations with Retrieval Augmented Generation: This article discusses a new technique for reducing AI-generated inaccuracies by augmenting large language models (LLMs) with proprietary data, which shows promise in enhancing the models’ knowledge base. #AI #LLMs #DataAugmentation #MachineLearning #TechnologyInnovation

AI Hallucinations : Fear Not — It’s A Solved Problem — Here’s How (With Examples!): The article discusses strategies to mitigate AI hallucinations in generative models, emphasising the necessity of integrating anti-hallucination measures across the entire Retrieval Augmented Generation (RAG) pipeline. It argues that achieving near-perfect control over hallucinations is crucial for reliability, drawing parallels to business standards in security and uptime. Techniques include thorough testing, leveraging economies of scale in SaaS platforms, and applying specific technical solutions like query pre-processing and dynamic context boundary walls in prompts. #AI #GenerativeModels #MachineLearning #AIethics #TechInnovation

EmoSpeaker: One-shot Fine-grained Emotion-Controlled Talking Face Generation: EmoSpeaker introduces a revolutionary technique for generating emotional talking-head videos from a single image, input audio, and specified emotion, capable of adjusting emotional intensity through fine-grained control. This method surpasses existing technologies in expression variation and lip-sync accuracy. #EmoSpeaker #TalkingHead #EmotionalVideo #TechInnovation #AIGeneratedContent

Sam Altman Seeks Trillions of Dollars to Reshape Business of Chips and AI: Sam Altman, CEO of OpenAI, is seeking to raise trillions to expand global semiconductor capabilities, aiming to address the shortage of AI chips and advance the development of artificial general intelligence. A trillion here, a trillion there. Pretty soon you’re talking real money. #SamAltman #OpenAI #SemiconductorIndustry #ArtificialIntelligence #TechInvestment

Machine Learning Research at Apple: What Apple does with machine intelligence in 2024 is anyone’s guess. Whereas the other Big Tech vendors tend to release incrementally, Apple (traditionally) likes to save up releases for one big announcement each year, so we will have to wait and see. Some breadcrumbs are starting to emerge. #MachineLearning #Apple #TechInnovation #AnnualRelease #BigTech

TikTok presents Boximator TikTok introduces Boximator, a tool for creating detailed and customisable motion in image-to-video transformations using box constraints and motion paths, exemplified by a girl in red covering her face with a skull, showcased through 10 unique examples. #Boximator #TikTokInnovation #ImageToVideo #CreativeTech #MotionGeneration

Cory Doctorow: What Kind of Bubble is AI?: Cory Doctorow’s article in Locus Magazine explores the nature of AI as a bubble, comparing it to previous tech bubbles. He discusses this bubble’s potential outcomes and remnants, highlighting the distinction between bubbles that leave valuable assets behind and those that do not. Doctorow expresses scepticism about AI’s sustainable value and business models, questioning what will remain when the hype subsides. #AIBubble #TechBubble #CoryDoctorow #FutureOfAI #TechScepticism

AI is the average of the Internet: WPP, Don’t become an AI North Korea: This strongly worded opinion piece from Punks and Pinstripes argues against WPP’s heavy investment in generative AI, likening it to a decoy masking stagnation akin to North Korea’s strategy with nuclear investment. It suggests that while AI can handle operational tasks efficiently, it stifles creativity in fields that thrive on human ingenuity, urging companies to balance AI use to avoid creative atrophy. #AIInvestment #CreativeAtrophy #BusinessStrategy #InnovationVsTradition #WPP

AI assistance is leading to lower code quality, claim researchers: Research suggests that while popular and enhancing productivity, AI coding assistants like GitHub’s Copilot may lead to lower code quality, with issues like increased code churn and higher amounts of repeated code. #AICoding #CodeQuality #DeveloperTools #TechResearch #SoftwareDevelopment

GPT4’s system prompt was leaked: This video breaks down the leaked GPT4 system prompt. The capabilities hinted at within the prompt are very surprising. For example, the policy statements for the use of DALL-E are particularly interesting with respect to emulating the style of artists. #GPT4 #SystemPrompt #AI #LLM #PromptEscape

Ancient Herculaneum scroll piece revealed by AI: Artificial intelligence has unlocked the contents of a papyrus scroll from Herculaneum, revealing a Greek philosopher’s insights on pleasure, previously hidden by the eruption of Mount Vesuvius 2000 years ago. This breakthrough, winning a $700,000 prize, could lead to more ancient texts being deciphered. #AncientTexts #AI #HerculaneumScrolls #Philosophy #VesuviusChallenge

Beyond Self-Attention: How a Small Language Model Predicts the Next Token: This article explores how a small transformer language model predicts the next token, focusing on the role of transformer blocks and feed-forward networks beyond multi-head self-attention. The author shares findings from a six-month investigation, proposing that each transformer block predicts the next tokens based on learned associations with classes of strings from the training data. #AI #MachineLearning #Transformers #LanguageModels #DeepLearning


[ED: If you’d like to sign up for this content as an email, click here to join the mailing list.]