AI Facts of 2024: A mid year Review 🤯
What's happened with AI in 2024? We are only half-way. It's blowing up! Here are the facts on what's happened so far, distilled from Dr Alan D. Thompson's report https://www.youtube.com/@DrAlanDThompson
Tara and Claude Sonnet 3.5
6/22/20242 min read
Taking a look back on the first half of 2024, it's clear to see that AI has blown-up. This article highlights some of the major facts about AI progress in 2024.
These 14 facts are taken from a review by Dr. Alan D. Thompson, a world-renowned expert in artificial intelligence, with a focus on augmenting human intelligence and advancing 'integrated AI'. As a former chairman for Mensa International's gifted families and an AI consultant to Fortune 500 companies, major governments, and UN member states, Dr. Thompson's insights are particularly valuable for small and medium-sized business owners looking to navigate the rapidly evolving AI landscape.
AI Adoption and Market Trends
1. 2024 is the year of AI adoption in enterprises, as predicted by OpenAI's COO.
2. 92% of Fortune 500 companies now subscribe to GPT services.
3. The average AI spend for Fortune 500 companies has reached $18 million per year.
AI Capabilities and Output
4. ChatGPT is outputting 100 billion words per day (70 million words per minute).
5. Over 120 AI models were released in the first half of 2024 alone.
6. Major AI labs have expanded beyond text data into multimodal data (text, vision, sound).
AI Performance Benchmarks
7. On the Google Proof Q&A (GPQA) benchmark:
- PhD-level experts achieved a score of 34%
- Domain experts with PhDs score about 65% on average
- Top AI models are outperforming human experts:
- GPT-4.0 achieves 53.6%
- Claude 3 Opus leads with 59.5%
Noteworthy AI Integrations
8. Deloitte onboarded 75,000 people into ChatGPT.
9. PwC onboarded 100,000 staff into GPT-4.
Future Implications
10. AI is approaching or surpassing human-level performance in various cognitive tasks.
11. AI is expected to tackle major challenges in education, health, economics, and scientific research.
Market Competition
12. Chinese AI models (e.g., GLM-4, Sova-5, Qwen-2) are now comparable to Silicon Valley's output.
13. More than half of all LLM releases in January 2024 were from China, indicating increased global competition.
Hardware Developments
14. NVIDIA is projected to sell 3.5 million H100 GPUs by the end of 2024, indicating massive infrastructure investments in AI.
Why does this matter?
In my opinion, this is a snapshot revealing where AI is positioned in the global economy and how it's integrating and filtering from the world's elite into the hands of the average joe. While enterprises have been able to capitalise on AI technology, SMBs are still yet to see the full advantages that are relatively accessible now.