Large Language Models and Commodity AI

The essential report for IT decision-makers
Edition #1

Large Language Models (LLM)
and Commodity AI

Large Language Models like GPT-3 and GPT-4 are fundamentally changing the way we approach natural language processing. They offer completely new possibilities for businesses across all industries and represent a technological breakthrough that happens “once every decade.” Additionally, this technology provides convenient API access. Alternatively, there are numerous open-source models that can be used locally and on-premise. Training your own model hardly seems worth it anymore. Everything you need to integrate AI into your products is available.

We call this Commodity AI.

In this report, we’ll show how your business can leverage the opportunities of Commodity AI to gain competitive advantages. You’ll get an overview of feasible use-case scenarios across various industries.

Ich
bin
hier.

What you get

  • Executive Summary
  • Industry analyses on use cases and opportunities
  • Decision finder: do it yourself or rely on APIs?
  • Design and development of AI-driven products
  • On-premises deployment of LLMs

Who’s writing here?

Dr. Larysa Visengeriyeva
INNOQ, Author ml-ops.org
Isabel Bär
INNOQ Alumna
Christoph Burnicki
INNOQ
Marcel Weiß
Analyst for the digital economy, neunetz.com

‍

We assist you

Let’s talk about your project, even if it’s just initial, rough thoughts. We’ll accompany you from ideation to implementation.

Identification of use cases In interactive workshops and trainings, we collaboratively determine suitable use cases for your company.

Architecture and technology consulting Together, we develop a sustainable Al strategy that doesn’t leave your existing systems, processes, and operations out.

Full Implementation by a team of multidisciplinary professionals.

References and further links

  1. Attention is all you need
  2. A good overview is given here: On the Opportunities and Risks of Foundation Models, chapter 1
  3. OpenAI’s massive GPT-3 model is impressive, but size isn’t everything
  4. Mosaic LLMs (Part 2): GPT-3 quality for <$500k
  5. Amazon EC2 Trn1 Instances
  6. On the Power of Foundation Models
  7. Model Summary table and region availability
  8. Growth of AI Through a Cloud Lens
  9. Über 600 Use Cases zur GPT-3-Nutzung
  10. A Brief Note To Our Founders Re: Impact Of Artificial Intelligence
  11. How will Language Modelers like ChatGPT Affect Occupations and Industries?
  12. ChatGPT, Generative AI and GPT-3 Apps and use cases
  13. Do the Rewards Justify the Means? Measuring Trade-Offs Between Rewards and Ethical Behavior in the MACHIAVELLI Benchmark
    ChatGPT-4 Outperforms Experts and Crowd Workers in Annotating Political Twitter Messages with Zero-Shot Learning:
  14. An Overview of the End-to-End Machine Learning Workflow
  15. AI Tasks: Paper. HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in HuggingFace
  16. Adapted of Data-centric Foundation Model Development: Bridging the gap between foundation models and enterprise AI
  17. Foundation Model Ops: Powering the Next Wave of Generative AI Apps
  18. Introducing LLaMA: A foundational, 65-billion-parameter large language model
  19. GitHub: llama.cpp
  20. Free Dolly: Introducing the World's First Truly Open Instruction-Tuned LLM
  21. Stability AI Launches the First of its StableLM Suite of Language Models
  22. LLaMA-Nachbau: RedPajama – erste dezentrale Open-Source-KI mit offenem Datensatz
  23. Meet Adobe Firefly
  24. Shop smarter with Shop.ai! We've brought our ChatGPT-powered shopping assistant to the web
  25. A European approach to artificial intelligence
  26. MLOps und Model Governance
  27. AI regulation: a pro-innovation approach
  28. OpenAI: ChatGPT back in Italy after meeting watchdog demands
  29. Responsible AI Belongs on the CEO Agenda
  30. On the Opportunities and Risks of Foundation Models
  31. More on the distribution of responsibilities along the AI value chain: 1 and 2
  32. Fairness-Aware Instrumentation of Preprocessing Pipelines for Machine Learning
  33. On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?
  34. Man is to Computer Programmer as Woman is to Homemaker? Debiasing Word Embeddings
  35. Definitionen für Fairness und Ableitungen von Teststrategien
  36. AI as a Target and Tool: An Attacker’s Perspective on ML
  37. Poisoning attacks on Machine Learning
    Dataset Security for Machine Learning: Data Poisoning, Backdoor Attacks, and Defenses
  38. Adversarial ML Threat Matrix
  39. Machine Learning Security – Teil 1
    Teil 2
  40. On the Opportunities and Risks of Foundation Models
  41. On the Opportunities and Risks of Foundation Models
  42. Monitoring Text-Based Generative AI Models Using Metrics Like Bleu Score
    Definitionen für Fairness und Ableitungen von Teststrategien
  43. Extrinsic and intrinsic evaluation and their purposes are described here: On the Opportunities and Risks of Foundation Models
    An in-depth article on how accountability can be achieved through a Model Governance Framework: MLOps und Model Governance
    Microsoft offers a Impact-Assessment-Template, that can be used for evaluation and documentation purposes. Google published a Whitepaper that mentions documentation obligations as part of an MLOps framework.
Ein Team aus 4 Personen, die gemeinsam an einem Tisch sitzen und diskutieren. Auf dem Tisch liegt ein Konzeptentwurf.
INNOQ

Our consultants have been advising SMEs & corporations for over 20 years, implementing IT systems of all sizes.

Our expertise is drawn from extensive hands-on experience in software architecture and development, platform operation and infrastructure, as well as digital product development.

We don’t view technology as an end in itself, but as an enabler for solving real-world problems.