Upgrade to Pro — share decks privately, control downloads, hide ads and more …

The Builder’s future in the generative AI era

The Builder’s future in the generative AI era

Mohammed Fazalullah

October 31, 2023
Tweet

More Decks by Mohammed Fazalullah

Other Decks in Technology

Transcript

  1. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Mohammed Fazalullah Qudrath Sr Developer Advocate, MENAT AWS The Builder’s future in the generative AI era
  2. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Agenda • Life before generative AI • Road to generative AI • How to equip yourself 2
  3. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. In simpler times… Proprietary software systems LAMP Reference books, CHMs
  4. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. 4
  5. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. 5
  6. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. What is Generative AI? 6
  7. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 7 Word Embeddings dog cat car bus
  8. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 8 Word Embeddings dog cat car bus person driver animal vehicle walk drive traffic tea
  9. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 9 Word Embeddings Language Models
  10. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 10 Word Embeddings Language Models Word2Vec
  11. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 11 Word Embeddings Text summarization Question answering Sentiment analysis Speech Language Understanding Language Models Natural Language Processing (NLP) Word2Vec
  12. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 12 The quick brown fox jumps over the lazy _ person. Next word prediction with simple word embedding. ? ? ? ? ? ? ? Sentence:
  13. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 13 The quick brown fox jumps over the lazy _ rabbit. Next word prediction with ‘memory’ such as Recurrent Neural Networks (RNN). ? ? ? Sentence:
  14. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Transformers - Attention Is All You Need - 2017 14 “Attention is all you need” Vaswani, Ashish & Shazeer, Noam & Parmar, Niki & Uszkoreit, Jakob & Jones, Llion & Gomez, Aidan & Kaiser, Lukasz & Polosukhin, Illia. (2017)
  15. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Transformers - Attention Is All You Need - 2017 15 “Attention is all you need” Vaswani, Ashish & Shazeer, Noam & Parmar, Niki & Uszkoreit, Jakob & Jones, Llion & Gomez, Aidan & Kaiser, Lukasz & Polosukhin, Illia. (2017) The Transformer model architecture. • “Self-attention” enables models to scale the understanding of relationships between words. • Efficiently use parallel computing.
  16. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 16 The quick brown fox jumps over the lazy _ dog.✓ Next word prediction with transformer based networks. Sentence:
  17. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The road to generative AI 17 The quick brown fox jumps over the lazy dog. Is an English-language pangram — a sentence that contains all the letters of the alphabet. The phrase is commonly used for touch-typing practice. Next word prediction with transformer based networks. Typing Tips Large Language Models (LLM) ✓✓✓
  18. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Rise of the machines 18 Subjective Performance Parameter Count (Model Size) GPT-2 1.5 billion GPT-3 175 billion BLO O M 176 billion BERT 345 m illion ?
  19. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Rise of the machines 19 Subjective Performance Parameter Count (Model Size) ? Foundation Models GPT-2 1.5 billion GPT-3 175 billion BLO O M 176 billion BERT 345 m illion
  20. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Applications for Generative AI 20 • Text generation (many types) • Game design • Industrial design • Drug design research • Image generation…
  21. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Text-to-Image 21 Language Models (Word embeddings) Image Models (Image embeddings) Text-to-Image Models (e.g. Stable Diffusion)
  22. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Prompt engineering 22 “Marigold the puppy on a space station”
  23. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Prompt engineering 23 Write a Blog How to win at Tic Tac Toe Title: Game theory Topic: 10 Length (paragraphs): OK
  24. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Prompt engineering 24 Write a Blog How to win at Tic Tac Toe Title: Game theory Topic: 10 Length (paragraphs): OK “You are a journalist for amazing-acme- blogs.com, an online publication for sophisticated game players aged 18 and up. Articles are written to inform and entertain and represent a unique and non- intuitive perspective. Write a blog post titled “How to win at Tic Tac Toe” on the topic of game theory. The article should be no longer than 10 paragraphs long. At the end list 3 prompts that can be used to generate images for the article.”
  25. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. chatbot = Chatbot( default_reply_length = 20, bot_persona = pmodel[‘friendly’], users_name =“User”, bots_name = “Chatbot”, topic = “tech support” ) prompt = """This is a friendly and safe chat session between a user and a computer called Chatbot.## User:I am a real person with a question to ask. Who are you?## Chatbot:I am a chatbot, and I am here to help.## User:What kind of questions will you answer?## Chatbot:I will answer questions about tech support.## User:My computer has crashed. What should I do?## Chatbot:""" A new way to program… 25
  26. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Generative AI on AWS: What you can use? 26
  27. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The easiest way to build and scale generative AI applications with foundation models (FMs) Amazon Bedrock Accelerate development of generative AI applications using FMs through an API, without managing infrastructure Choose FMs from Amazon, AI21 Labs, Anthropic, Cohere, Meta, and Stability AI to find the right FM for your use case Privately customize FMs using your organization’s data G E N E R A L L Y A V A I L A B L E
  28. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. JURASSIC-2 Multilingual LLMs for text generation in Spanish, French, German, Portuguese, Italian, and Dutch CLAUDE 2 LLM for conversations, question answering, and workflow automation based on research into training honest and responsible AI systems Generation of unique, realistic, high-quality images, art, logos, and designs SDXL 1.0 Text summarization, generation, classification, open- ended Q&A, information extraction, embeddings and search AMAZON TITAN COMMAND Text generation model for business applications like summarization, copywriting, dialog, extraction, and question answering Pre-trained and fine-tuned LLMs for natural language tasks like question answering and reading comprehension LLAMA 2 Amazon Bedrock: Foundation Models C O M I N G S O O N P R E V I E W P R E V I E W
  29. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Amazon CodeWhisperer: AI-based Code Companion 29
  30. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Where to from here? • Models will keep getting bigger, and more sophisticated. • Many developers will use foundation models and prompt engineering. • Training costs and sustainability come into focus. • Inference costs and sustainability come into focus. 30 • Algorithm optimization. • Dedicated hardware, designed for LLM, such as AWS Trainium reduce costs. • Model pruning and other optimization. • Dedicated hardware, designed for LLM, such as AWS Inferentia reduce costs.
  31. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. How to equip yourself 31
  32. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The promise for developers • Productivity increase with boilerplate code generation, bug fixes, documentation from prompts • Democratize coding, lower barrier to entry • New development workflows • Leverage prompt engineering as the new way of “rubber duck” debugging 32
  33. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Impacts • Impacts on testing • Impacts on security • Impacts on operations 33
  34. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Staying relevant as a developer • Develop prompt engineering skills • Focus on high-value creative tasks • Develop skills in directing/pruning AI output • Continuous learning mindset • The future is collaborative 34
  35. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Generative AI through the lens of Dunning-Kruger effect 35 1 2 3 4 5 1. Wow! 2. Huh? 3. Does this even work? 4. Ah, so that’s how it works? 5. I think I know how to use it better! No knowledge (believe the hype) Enough knowledge (understand reality) Knowledge Confidence level 0% 100%
  36. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. © 2023, Amazon Web Services, Inc. or its affiliates. Generative AI on AWS: What you can use? 36
  37. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Amazon SageMaker JumpStart: Foundation Models 37 Amazon SageMaker JumpStart Access and try out public and proprietary FMs, and customize and integrate them into your generative AI applications Deploy Deploy the model and run inference for your generative AI use case Browse Browse public and proprietary FMs Experiment Experiment with FMs before choosing a model for deployment Customize Customize selected FM with your own dataset without training from scratch
  38. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. A V A I L A B L E O N S A G E M A K E R J U M P S T A R T Models Stable Diffusion XL 1.0 2.1 base Upscaling Inpainting Tasks Generate photo-realistic images from text input Improve quality of generated images Features Fine-tuning on Stable Diffusion 2.1 base model Models AlexaTM 20B Tasks Machine translation Question answering Summarization Annotation Data generation Models Falcon-7B, 40B Open LlaMA RedPajama MPT-7B BloomZ 176B Flan T-5 models (8 variants) DistilGPT2 GPT NeoXT Bloom models (3 variants) Tasks Machine translation Question answering Summarization Models Cohere Command XL Tasks Text generation Information extraction Question answering Summarization Models Jurassic-2 Ultra, Mid Contextual answers Summarize Paraphrase Grammatical error correction Tasks Text generation Long-form generation Summarization Paraphrasing Chat Information extraction Models Lyra-Fr 10B, Mini Tasks Text generation Keyword extraction Information extraction Question answering Summarization Sentiment analysis Classification Models Llama 2 7B, 13B, 70B Tasks Question answering Chat Summarization Paraphrasing Sentiment analysis Text generation Models Dolly Tasks Question answering Chat Summarization Paraphrasing Sentiment analysis Text generation Amazon SageMaker JumpStart: Foundation Models
  39. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. The easiest way to build and scale generative AI applications with foundation models (FMs) Amazon Bedrock Accelerate development of generative AI applications using FMs through an API, without managing infrastructure Choose FMs from Amazon, AI21 Labs, Anthropic, Cohere, Meta, and Stability AI to find the right FM for your use case Privately customize FMs using your organization’s data G E N E R A L L Y A V A I L A B L E
  40. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. JURASSIC-2 Multilingual LLMs for text generation in Spanish, French, German, Portuguese, Italian, and Dutch CLAUDE 2 LLM for conversations, question answering, and workflow automation based on research into training honest and responsible AI systems Generation of unique, realistic, high-quality images, art, logos, and designs SDXL 1.0 Text summarization, generation, classification, open- ended Q&A, information extraction, embeddings and search AMAZON TITAN COMMAND Text generation model for business applications like summarization, copywriting, dialog, extraction, and question answering Pre-trained and fine-tuned LLMs for natural language tasks like question answering and reading comprehension LLAMA 2 Amazon Bedrock: Foundation Models C O M I N G S O O N P R E V I E W P R E V I E W
  41. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Trn1n/Inf2: Purpose-built accelerators for generative AI AWS Inferentia Lowest cost per inference in the cloud for running deep learning (DL) models Up to 70% lower cost per inference than comparable Amazon EC2 instances AWS Trainium The most cost-efficient, high- performance training of LLMs and diffusion models Up to 50% savings on training costs over comparable Amazon EC2 instances AWS Inferentia2 High performance at the lowest cost per inference for LLMs and diffusion models Up to 40% better price performance than comparable Amazon EC2 instances 41
  42. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Amazon CodeWhisperer: AI-based Code Companion 42
  43. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. © 2023, Amazon Web Services, Inc. or its affiliates. Demo! 43
  44. THE BUILDERS FUTURE IN THE GENERATIVE AI ERA © 2024,

    Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. © 2024, Amazon Web Services, Inc. or its affiliates. All rights reserved. Amazon Confidential and Trademark. Thank you! Mohammed Fazalullah Qudrath Sr Developer Advocate, MENAT AWS