Artificial Intelligence

5 AI projects that (actually) get you hired. Most resumes get ignored, these won't: 1 → RAG from Scratch Build retrieval systems properly. No framework shortcuts. https://github.com/langchain-ai/rag-from-scratch 2 → AI Social Media Agent Autonomous content generation. Real world automation. https://github.com/langchain-ai/social-media-agent 3 → Medical Image Analysis Healthcare AI applications. Production ready pipeline. https://github.com/databricks-industry-solutions/pixels 4 → MCP Tool Calling Agents Multi tool orchestration. Agent architecture mastery. https://docs.databricks.com/aws/en/notebooks/source/generative-ai/langgraph-mcp-tool-calling-agent.html 5 → AI Assistant Memory Persistent conversation systems. Context management solved. https://lnkd.in/gnA2Xmzw These prove you can ship. Not just learn.

This is huge. Now you can use Claude Code for FREE: Ollama is now compatible with the anthropic messages API. which means you can use Claude code with open-source models. Think about that for a second. the entire Claude harness: - the agentic loops - the tool use - the coding workflows All powered by private LLMs running on your own machine. https://dailydoseofds.github.io/ai-engg-book?trk=public_post_comment-text

🚨 BIG news for students! 🚨 College students can now get 1 YEAR FREE of Microsoft 365 Premium - AI + LinkedIn Premium + 🎓💻 That means: ✨ Career tools on LinkedIn ✨ Get the ultimate AI experience ✨ Word, Excel, PowerPoint & more ✨ Resume building, job prep, and productivity — all free This is one of the most exciting student perks Microsoft launched 🙌 Don’t miss it —share with every college student you know! 🔗 Link: https://www.microsoft.com/en-us/microsoft-365/college-student-pricing

The most expensive AI education in the world is now FREE — most will ignore it 🛑 That’s the real gap in 2026. Next year, winners won’t be the people who know AI. They’ll be the ones who turn complexity into progress while others stay busy and burnt out. After leading AI and digital transformation across legal tech, housing, government, and professional bodies — here are the 10 capabilities that actually move careers and companies forward: 1️⃣ Prompt Engineering Clarity beats cleverness — context, constraints, examples create repeatable quality. 2️⃣ AI Workflow Automation Friction is the enemy — automate invisible work to reclaim strategic bandwidth. 3️⃣ AI Agents Outcomes > tasks — agents connect intent to results and behave like teammates. 4️⃣ RAG (Retrieval-Augmented Generation) Your answers already exist — unlock siloed knowledge instantly. 5️⃣ Multimodal AI More context, fewer errors — text + visuals + voice changes understanding. 6️⃣ Domain-Specific Assistants Bigger models don’t win — models that think like your business do. 7️⃣ Voice AI & Avatars Explain once, scale forever — onboarding and training without repetition. 8️⃣ AI Tool Stacking No single tool wins — the right stack breaks bottlenecks. 9️⃣ AI Video Generation Speed builds trust — iterate fast, test often, improve weekly. 🔟 LLM Management Control matters — track cost, latency, and performance as usage scales. Unpopular opinion: Don’t chase tools — build systems that compound impact. AI’s value isn’t intelligence, It’s leverage.

The most expensive AI education in the world is now FREE — most will ignore it 🛑 That’s the real gap in 2026. Next year, winners won’t be the people who know AI. They’ll be the ones who turn complexity into progress while others stay busy and burnt out. After leading AI and digital transformation across legal tech, housing, government, and professional bodies — here are the 10 capabilities that actually move careers and companies forward: 1️⃣ Prompt Engineering Clarity beats cleverness — context, constraints, examples create repeatable quality. 2️⃣ AI Workflow Automation Friction is the enemy — automate invisible work to reclaim strategic bandwidth. 3️⃣ AI Agents Outcomes > tasks — agents connect intent to results and behave like teammates. 4️⃣ RAG (Retrieval-Augmented Generation) Your answers already exist — unlock siloed knowledge instantly. 5️⃣ Multimodal AI More context, fewer errors — text + visuals + voice changes understanding. 6️⃣ Domain-Specific Assistants Bigger models don’t win — models that think like your business do. 7️⃣ Voice AI & Avatars Explain once, scale forever — onboarding and training without repetition. 8️⃣ AI Tool Stacking No single tool wins — the right stack breaks bottlenecks. 9️⃣ AI Video Generation Speed builds trust — iterate fast, test often, improve weekly. 🔟 LLM Management Control matters — track cost, latency, and performance as usage scales. Unpopular opinion: Don’t chase tools — build systems that compound impact. AI’s value isn’t intelligence, It’s leverage.

🚀 If you’re entering an AI career right now, here’s the truth: It’s not about learning “everything.” It’s about learning the right technical foundations — the ones the industry actually uses. These are the core skills that will matter for the next 5–10 years, no matter how fast AI evolves 👇 1️⃣ Learn how modern LLMs actually work You don’t need to know the math behind transformers, but you must understand: • tokens & embeddings • context windows • attention • prompting vs reasoning • fine-tuning vs RAG • when models hallucinate (and why) If you don’t know how the engine works, you can’t drive it well. 2️⃣ Learn Retrieval — the real backbone of enterprise AI Most AI applications in companies rely on RAG, not fine-tuning. Focus on: • chunking strategies • embedding models • hybrid retrieval (dense + sparse) • vector databases • knowledge graphs • context filtering • evaluation of retrieved docs If you master retrieval, you instantly become valuable. 3️⃣ Learn how to evaluate AI systems, not just build them Engineers build models. Professionals who can evaluate them are the ones who get promoted. Learn to measure: • grounding accuracy • relevance • completeness • tool-use correctness • consistency across runs • latency • safety This is where the real skill gap is. 4️⃣ Learn prompting as an engineering discipline Not “try random prompts.” But systematic methods like: • template prompts • tool-calling prompts • guardrail prompts • chain-of-thought • reflection prompts • constraint-based prompting Prompting is becoming the new API design. 5️⃣ Learn how to build agentic workflows AI is moving from answers → decisions → actions. You should know: • planner → executor → verifier agent structure • tool routing • action space design • human-in-the-loop workflows • permissioning • error recovery loops This is what separates beginners from real AI engineers. 6️⃣ Learn Python + APIs deeply You don’t need to be a software engineer, but you must be comfortable with: • Python basics • API calls • JSON • LangChain / LlamaIndex / DSPy • building small scripts • reading logs • debugging AI pipelines This is the “plumbing” behind AI systems. 7️⃣ Build real projects, not toy demos Instead of “build a chatbot,” build: • a support email classifier • a RAG system on company policies • a customer insights extractor • an automatic meeting summarizer • a multimodal analyzer (text + image) • an internal tool-calling agent Projects that solve real problems get you hired. 8️⃣ Learn one domain deeply AI generalists struggle. AI + domain experts win. Choose one: • finance • healthcare • retail • manufacturing • real estate • cybersecurity • operations • supply chain • HR tech AI skill + domain depth = career acceleration. If you’re entering AI today: Focus on retrieval, reasoning, evaluation, agents, and real projects. These are the skills companies are desperate for.

From AI Prototype to Production — One Guide That Matters If you’re building AI agents and wondering how to take them from demo to real-world deployment, this Google Cloud whitepaper is gold. It explains, in simple terms: • How to deploy AI agents safely • How to scale them for enterprise use • CI/CD, observability & trust in production • Real challenges of moving from prototype → production • Agent-to-Agent (A2A) interoperability Perfect for AI/ML engineers, DevOps teams, and architects working on serious AI systems. 📄 Read here: https://www.kaggle.com/whitepaper-prototype-to-production Sharing this because production-ready AI is where real value is created 💡

We’re building a science-first AI team at 1st Principles AI Labs. Looking for CS grads (1–3 yrs) with: ⚙️ Strong Python (core + scientific coding) 📐 Solid CS theory + math ⚡️ High energy + problem-solving mindset No AI/ML experience needed. We’ll teach you from first principles to frontier-level AI. https://1stprinciples.ai/careers

So the most important question as one evaluates the frenzied AI investment landscape is not really whether it will pop or not, but what sort of legacy it will leave behind. Would the fallout include a hobbled financial system and an intractable, prolonged recession, as the bursting of the housing bubble left in its wake? Or is it more likely to look like the dot-com bubble, whose bursting produced a comparatively shallow economic downturn and ultimately gave the world the modern internet? As I pointed out in my last column about AI, Gita Gopinath, former chief economist of the International Monetary Fund, calculated that a stock market crash equivalent to that which ended the dot-com boom would erase some $20tn in American household wealth and another $15tn abroad, enough to strangle consumer spending and induce a recession. But the economic pain would depend to a large extent on how the AI investment surge is being financed. One problem is that we don’t really know. The housing bubble was built from a boom in mortgage finance, as yield-seeking banks stuffed themselves with bonds built of bundles of mortgages to increasingly uncreditworthy borrowers. When the borrowers couldn’t pay, the boom left a forest of damaged balance sheets in its wake, from over-indebted households with no access to credit, to a banking system hobbled by worthless bonds. Financing froze. It took years for America’s credit-driven economy to recover. AI could produce a similar landscape. A critical determinant is how much debt is at stake. It wouldn’t be such a problem if the bubble were financed largely from the cash pile of Alphabet and Amazon, Microsoft and Facebook. They might lose their shirt, but who cares. The worrying bit is that it seems they are increasingly relying on borrowing, which means the prospect of a bursting bubble would again put the financial system at risk. Big Tech has raised nearly $250bn in debt so far this year, according to Bloomberg, a record. Analysts at Morgan Stanley suggest that debt will be needed to fill a $1.5tn funding gap to ramp up spending on data centers and hardware. Problematically, it is getting hard to follow the money, as Nvidia, Open AI and others in the ecosystem buy into each other, clouding who, in the end, will be left holding the bag. The other question is to what extent the AI that the Silicon Valley faithful are building will endure. Railways survived the 19th century railway bust. The Internet survived the dot-com implosion. Is there anything of sufficient value to justify the current moment of euphoria, even if it heads south for a time? Until a few weeks ago, I would have said sure: there must be something in Chat GPT or Claude that will raise business productivity. But to justify the vast quantities of money they are going to have to build something really impressive – as in superhuman general intelligence impressive. Over the last several weeks, a thought has bubbled up through the ecosystem that they won’t. It’s a thought built on the thoughts of techier minds than mine. Yann LeCun, until recently Meta’s chief scientist and a winner of the Turing Award, has been saying that the massive spend on Large Language Models that today define the AI space is misguided. Artificial General Intelligence – aka the Superhuman – can only come about by dropping LLMs – which are essentially massive correlation engines – and switching to something else called a world model architecture, where machines develop a “mental” model of the outside world. If he’s right, that would be one big oops for much of today’s AI spend. Nvidia and the rest of us may be about to learn, once again, that just because you sold a load of jeans and shovels, it doesn’t mean there is gold in them thar hills. - Eduardo Porter, The Guardian

The question isn’t whether the AI bubble will burst – but what the fallout will be!! The California Gold Rush left an outsized imprint on America. Some 300,000 people flocked there from 1848 to 1855, from as far away as the Ottoman Empire. Prospectors massacred Indigenous people to take the gold from their lands in the Sierra Nevada mountains. And they boosted the economies of nearby states and faraway countries from whence they bought their supplies. Gold provided the motivation for California – a former Mexican territory then controlled by the US military – to become a state with laws of its own. And yet, few “49ers” as prospectors were known, struck it rich. It was the merchants selling prospectors food and shovels who made the money. One, a Bavarian immigrant named Levi Strauss who sold denim overalls to the gold bugs passing through San Francisco, may be the most remembered figure of his day. California is going through another investment rush these days. This time it’s centered in Silicon Valley. The pot of gold is more elusive but potentially much bigger: Artificial Intelligence. What this rush leaves in its wake will shape the long-term future of civilization – or maybe not? The question everyone seems to be asking is: is AI a bubble? Lots of people seem to think so, including Open AI’s Sam Altman and the Bank of England. How else to explain Nvidia’s stock price, which more than doubled from April to November, based entirely on the expectation, nay hope, that AI will produce a super-intelligence that can do everything humans do but better. Nvidia – like Levi Strauss back in the day – is at least selling something: computer chips. The valuations of many of the other AI plays – like Open AI or Anthropic – are based largely on the dream. The big analytical challenge, however, is to figure out what kind of bubble this is. Is it the kind that will ravage the economy when it bursts? What will it leave of value once it pops? Bubbles all share one characteristic – besotted investors in pursuit of a dream. But they come in many flavors. Not 20 years ago, we suffered the housing bubble, when home prices rose to stratospheric heights and almost brought down the financial system as they crashed back to earth. Less than a decade earlier, it was the dot-com bubble that burst, when investors realized that Webvan, Pets.com and the like were not worth billions just because they used the Internet. A few years before that we witnessed the rise and collapse of the East Asian bubble – with ancillary bubblettes in Russia and Brazil – when money rushed into these emerging markets, freaked and rushed out. There was the Tequila Crisis, which pummeled the Mexican peso and its economy. And the Japanese bubble, when the value of the Nikkei 225 stock index tripled over four years before it fell by 60% over the next two and a half. Bubbles have plagued the world’s finances at least since the 17th century, when Dutch investors fell in and out of love with tulips. In the 18th century, French, Dutch and British investors produced what came to be known as the South Sea bubble by giving in to euphoria over the value of potential of new trade routes across the Atlantic. That bubble ended with an Act of the British Parliament “to Restrain the Extravagant and Unwarrantable Practice of Raising Money by Voluntary Subscription For Carrying on Projects Dangerous to the Trade and Subjects of the United Kingdom.” It came to be known as the Bubble Act. Virtually every new frontier opened up to investment has led to a speculative bubble. Investors have scrambled to tap into its promise only to overdo it and stampede in retreat. Economists Carmen Reinhart and Kenneth Rogoff found that of the world’s 66 major economies, including developed nations and big developing countries, only Portugal, Austria, Belgium and the Netherlands had avoided a banking crisis between 1945 and 2007. By the end of 2008 none of them were unscathed. https://youtu.be/wuB4UHj67Xs?si=yNsT551HcR_Yiq4c

So the most important question as one evaluates the frenzied AI investment landscape is not really whether it will pop or not, but what sort of legacy it will leave behind. Would the fallout include a hobbled financial system and an intractable, prolonged recession, as the bursting of the housing bubble left in its wake? Or is it more likely to look like the dot-com bubble, whose bursting produced a comparatively shallow economic downturn and ultimately gave the world the modern internet? As I pointed out in my last column about AI, Gita Gopinath, former chief economist of the International Monetary Fund, calculated that a stock market crash equivalent to that which ended the dot-com boom would erase some $20tn in American household wealth and another $15tn abroad, enough to strangle consumer spending and induce a recession. But the economic pain would depend to a large extent on how the AI investment surge is being financed. One problem is that we don’t really know. The housing bubble was built from a boom in mortgage finance, as yield-seeking banks stuffed themselves with bonds built of bundles of mortgages to increasingly uncreditworthy borrowers. When the borrowers couldn’t pay, the boom left a forest of damaged balance sheets in its wake, from over-indebted households with no access to credit, to a banking system hobbled by worthless bonds. Financing froze. It took years for America’s credit-driven economy to recover. AI could produce a similar landscape. A critical determinant is how much debt is at stake. It wouldn’t be such a problem if the bubble were financed largely from the cash pile of Alphabet and Amazon, Microsoft and Facebook. They might lose their shirt, but who cares. The worrying bit is that it seems they are increasingly relying on borrowing, which means the prospect of a bursting bubble would again put the financial system at risk. Big Tech has raised nearly $250bn in debt so far this year, according to Bloomberg, a record. Analysts at Morgan Stanley suggest that debt will be needed to fill a $1.5tn funding gap to ramp up spending on data centers and hardware. Problematically, it is getting hard to follow the money, as Nvidia, Open AI and others in the ecosystem buy into each other, clouding who, in the end, will be left holding the bag. The other question is to what extent the AI that the Silicon Valley faithful are building will endure. Railways survived the 19th century railway bust. The Internet survived the dot-com implosion. Is there anything of sufficient value to justify the current moment of euphoria, even if it heads south for a time? Until a few weeks ago, I would have said sure: there must be something in Chat GPT or Claude that will raise business productivity. But to justify the vast quantities of money they are going to have to build something really impressive – as in superhuman general intelligence impressive. Over the last several weeks, a thought has bubbled up through the ecosystem that they won’t. It’s a thought built on the thoughts of techier minds than mine. Yann LeCun, until recently Meta’s chief scientist and a winner of the Turing Award, has been saying that the massive spend on Large Language Models that today define the AI space is misguided. Artificial General Intelligence – aka the Superhuman – can only come about by dropping LLMs – which are essentially massive correlation engines – and switching to something else called a world model architecture, where machines develop a “mental” model of the outside world. If he’s right, that would be one big oops for much of today’s AI spend. Nvidia and the rest of us may be about to learn, once again, that just because you sold a load of jeans and shovels, it doesn’t mean there is gold in them thar hills. - Eduardo Porter, The Guardian

The question isn’t whether the AI bubble will burst – but what the fallout will be The California Gold Rush left an outsized imprint on America. Some 300,000 people flocked there from 1848 to 1855, from as far away as the Ottoman Empire. Prospectors massacred Indigenous people to take the gold from their lands in the Sierra Nevada mountains. And they boosted the economies of nearby states and faraway countries from whence they bought their supplies. Gold provided the motivation for California – a former Mexican territory then controlled by the US military – to become a state with laws of its own. And yet, few “49ers” as prospectors were known, struck it rich. It was the merchants selling prospectors food and shovels who made the money. One, a Bavarian immigrant named Levi Strauss who sold denim overalls to the gold bugs passing through San Francisco, may be the most remembered figure of his day. California is going through another investment rush these days. This time it’s centered in Silicon Valley. The pot of gold is more elusive but potentially much bigger: Artificial Intelligence. What this rush leaves in its wake will shape the long-term future of civilization – or maybe not? The question everyone seems to be asking is: is AI a bubble? Lots of people seem to think so, including Open AI’s Sam Altman and the Bank of England. How else to explain Nvidia’s stock price, which more than doubled from April to November, based entirely on the expectation, nay hope, that AI will produce a super-intelligence that can do everything humans do but better. Nvidia – like Levi Strauss back in the day – is at least selling something: computer chips. The valuations of many of the other AI plays – like Open AI or Anthropic – are based largely on the dream. The big analytical challenge, however, is to figure out what kind of bubble this is. Is it the kind that will ravage the economy when it bursts? What will it leave of value once it pops? Bubbles all share one characteristic – besotted investors in pursuit of a dream. But they come in many flavors. Not 20 years ago, we suffered the housing bubble, when home prices rose to stratospheric heights and almost brought down the financial system as they crashed back to earth. Less than a decade earlier, it was the dot-com bubble that burst, when investors realized that Webvan, Pets.com and the like were not worth billions just because they used the Internet. A few years before that we witnessed the rise and collapse of the East Asian bubble – with ancillary bubblettes in Russia and Brazil – when money rushed into these emerging markets, freaked and rushed out. There was the Tequila Crisis, which pummeled the Mexican peso and its economy. And the Japanese bubble, when the value of the Nikkei 225 stock index tripled over four years before it fell by 60% over the next two and a half. Bubbles have plagued the world’s finances at least since the 17th century, when Dutch investors fell in and out of love with tulips. In the 18th century, French, Dutch and British investors produced what came to be known as the South Sea bubble by giving in to euphoria over the value of potential of new trade routes across the Atlantic. That bubble ended with an Act of the British Parliament “to Restrain the Extravagant and Unwarrantable Practice of Raising Money by Voluntary Subscription For Carrying on Projects Dangerous to the Trade and Subjects of the United Kingdom.” It came to be known as the Bubble Act. Virtually every new frontier opened up to investment has led to a speculative bubble. Investors have scrambled to tap into its promise only to overdo it and stampede in retreat. Economists Carmen Reinhart and Kenneth Rogoff found that of the world’s 66 major economies, including developed nations and big developing countries, only Portugal, Austria, Belgium and the Netherlands had avoided a banking crisis between 1945 and 2007. By the end of 2008 none of them were unscathed.

The best fine-tuning guide you'll find on arXiv this year. Covers: > NLP basics > PEFT/LoRA/QLoRA techniques > Mixture of Experts > Seven-stage fine-tuning pipeline Source: https://arxiv.org/pdf/2408.13296v1

🚗 If ML Algorithms Were Cars… 🚙 Linear Regression — Maruti 800 Simple, reliable, gets you from A to B. Struggles on curves, but hey… classic. 🚕 Logistic Regression — Auto-rickshaw Only two states: yes/no, 0/1, go/stop. Efficient, but not built for complex roads. 🚐 Decision Tree — Old School Jeep Takes sharp turns at every split. Fun, but flips easily. 😅 🚜 Random Forest — Tractor Convoy A lot of vehicles working together. Slow individually, powerful as a group. 🏎 SVM — Ferrari Elegant, fast, and only useful when the road (data) is perfectly separated. Otherwise… good luck. 🚘 KNN — School Bus Just follows the nearest kids and stops where they stop. Zero intelligence, full blind faith. 🚛 Naive Bayes — Delivery Van Simple, fast, predictable. Surprisingly efficient despite assumptions that make no sense. 🚗💨 Neural Network — Tesla Lots of hidden features, runs on massive power. Even mechanics (developers) can't fully explain how it works. 🚀 Deep Learning — SpaceX Rocket Needs crazy fuel, insane computing power, and one wrong parameter = explosion. But when it works… mind-blowing. 🏎💥 Gradient Boosting — Formula 1 Car Tiny improvements stacked until it becomes a monster. Warning: overheats (overfits) if not tuned properly. 🤖 Reinforcement Learning — Self-Driving Car Learns by trial and error. Sometimes brilliant… sometimes crashes into a wall.

Synthetic Image Detection using Gradient Fields 💡 A simple luminance-gradient PCA analysis reveals a consistent separation between real photographs and diffusion-generated images. Real images produce coherent gradient fields tied to physical lighting and sensor characteristics, while diffusion samples show unstable high-frequency structures from the denoising process. By converting RGB to luminance, computing spatial gradients, flattening them into a matrix, and evaluating the covariance through PCA, the difference becomes visible in a single projection. This provides a lightweight and interpretable way to assess image authenticity without relying on metadata or classifier models.

AI research is exploding 🔥— thousands of new papers every month. But these 9 built the foundation. Most developers jump straight into LLMs without understanding the foundational breakthroughs. Here's your reading roadmap ↓ 1️⃣ 𝐄𝐟𝐟𝐢𝐜𝐢𝐞𝐧𝐭 𝐄𝐬𝐭𝐢𝐦𝐚𝐭𝐢𝐨𝐧 𝐨𝐟 𝐖𝐨𝐫𝐝 𝐑𝐞𝐩𝐫𝐞𝐬𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧𝐬 𝐢𝐧 𝐕𝐞𝐜𝐭𝐨𝐫 𝐒𝐩𝐚𝐜𝐞 (𝟐𝟎𝟏𝟑) Where it all began. Introduced word2vec and semantic word understanding. → Made "king - man + woman = queen" math possible → 70K+ citations, still used everywhere today 🔗 https://arxiv.org/abs/1301.3781 2️⃣ 𝐀𝐭𝐭𝐞𝐧𝐭𝐢𝐨𝐧 𝐈𝐬 𝐀𝐥𝐥 𝐘𝐨𝐮 𝐍𝐞𝐞𝐝 (𝟐𝟎𝟏𝟕) Killed RNNs. Created the Transformer architecture. → Every major LLM uses this foundation 🔗 https://arxiv.org/pdf/1706.03762 3️⃣ 𝐁𝐄𝐑𝐓 (𝟐𝟎𝟏𝟖) Stepping stone on Transformer architecture. Introduced bidirectional pretraining for deep language understanding. → Looks left AND right to understand meaning 🔗 https://arxiv.org/pdf/1810.04805 4️⃣ 𝐆𝐏𝐓 (𝟐𝟎𝟏𝟖) Unsupervised pretraining + supervised fine-tuning. → Started the entire GPT revolution 🔗 https://cdn.openai.com/research-covers/language-unsupervised/language_understanding_paper.pdf 5️⃣ 𝐂𝐡𝐚𝐢𝐧-𝐨𝐟-𝐓𝐡𝐨𝐮𝐠𝐡𝐭 𝐏𝐫𝐨𝐦𝐩𝐭𝐢𝐧𝐠 (𝟐𝟎𝟐𝟐) "Think step by step" = 3x better reasoning 🔗 https://arxiv.org/pdf/2201.11903 6️⃣ 𝐒𝐜𝐚𝐥𝐢𝐧𝐠 𝐋𝐚𝐰𝐬 𝐟𝐨𝐫 𝐍𝐞𝐮𝐫𝐚𝐥 𝐋𝐚𝐧𝐠𝐮𝐚𝐠𝐞 𝐌𝐨𝐝𝐞𝐥𝐬 (𝟐𝟎𝟐𝟎) Math behind "bigger = better" → Predictable power laws guide AI investment 🔗 https://arxiv.org/pdf/2001.08361 7️⃣ 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠 𝐭𝐨 𝐒𝐮𝐦𝐦𝐚𝐫𝐢𝐳𝐞 𝐰𝐢𝐭𝐡 𝐇𝐮𝐦𝐚𝐧 𝐅𝐞𝐞𝐝𝐛𝐚𝐜𝐤 (𝟐𝟎𝟐𝟎) Introduced RLHF - the secret behind ChatGPT's helpfulness 🔗 https://arxiv.org/pdf/2009.01325 8️⃣ 𝐋𝐨𝐑𝐀 (𝟐𝟎𝟐𝟏) Fine-tune 175B models by training 0.01% of weights → Made LLM customization affordable for everyone 🔗 https://arxiv.org/pdf/2106.09685 9️⃣ 𝐑𝐞𝐭𝐫𝐢𝐞𝐯𝐚𝐥-𝐀𝐮𝐠𝐦𝐞𝐧𝐭𝐞𝐝 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐨𝐧 (𝟐𝟎𝟐𝟎) Original RAG paper - combines retrieval with generation → Foundation of every knowledge-grounded AI system 🔗 https://arxiv.org/abs/2005.11401