Three Important Tech Stories of 2024 – Asrar Qureshi’s Blog Post 1050

Three Important Tech Stories of 2024 – Asrar Qureshi’s Blog Post 1050

Dear Colleagues! This is Asrar Qureshi’s Blog Post 1050 for Pharma Veterans. Pharma Veterans Blogs are published by Asrar Qureshi on its dedicated site https://pharmaveterans.com. Please email to pharmaveterans2017@gmail.com for publishing our contributions here.

Credit: Tara Winstead

Credit: Darlene Anderson

Credit: Mizuno K

Credit: Moose Photos

Preamble

Year 2024 saw revolutionary changes in various areas of technology. Most notable were in the field of or related to Artificial Intelligence – AI. Many people are writing about technology every day or week, and mostly on AI. I read a few regularly and I have come to like a short daily newsletter from Martin Crowley and Liam Lawson, titled ‘AI Tool Report’. This blogpost is also based on it. Link at the end.

Top 3 Stories of 2024

#1 – Open AI Launches First Reasoning Model – o1

In September 2024, OpenAI launched its first reasoning models—o1-preview (for general users and o1 mini (for writing code)—which “spend more time thinking” before they respond, and can answer more complex questions, faster than a human can, marking a huge step forward in AI.

Unlike previous models, o1 has been trained with “reinforcement learning” which teaches it to “think” before answering by issuing rewards and penalties, which allows it to self-solve problems and fact-check answers.

Although they take a few seconds to ‘think’ and calibrate answers, tests demonstrated that the o1 models could accurately solve 83% of complex Math problems, whereas older versions could solve just 13%.

When answering scientific problems, o1 outperformed PhD-level scientists in problem-solving, which, although marks a huge AI advancement, has worried experts, who are now calling for better AI regulation.

The launch of OpenAI’s o1 has sparked the release of a flurry of reasoning models from Google (Gemini 2.0 Flash Thinking Experimental), DeepSeek (DeepSeek-R1), and Alibaba (QwQ-32B-Preview) which marks an important shift in AI’s ability to reason through problems and paves the way for a future where AI can work, like humans, to solve complex tasks.

Benefits  

Advanced Problem-Solving Capabilities: OpenAI’s o1 reasoning model excels at complex reasoning tasks, enabling users to tackle problems that require logical analysis and multifaceted decision-making.  

Natural Language Interaction: The model’s enhanced understanding of context and nuanced queries allows for clear, human-like communication and explanation of its reasoning processes.  

Time Efficiency: By quickly analyzing and solving problems, o1 reduces the time required for tasks like research, programming, and complex calculations.  

Diverse Applications: Industries like healthcare (e.g., diagnostics), education (personalized learning), and business (strategy development) can benefit significantly from o1’s reasoning power.  

Supports Creativity and Innovation: o1 assists in brainstorming, scenario analysis, and generating innovative ideas, making it a useful tool for creative professionals.  

Risks

Over-Reliance on AI: Users may depend too heavily on the model, leading to reduced critical thinking skills or overlooking its limitations in nuanced or subjective scenarios.  

Bias in Reasoning: If the training data contains biases, o1 might replicate or amplify those biases in its reasoning and solutions.  

Misinformation Risks: While o1’s reasoning is sophisticated, it could still generate plausible but incorrect answers, leading to potential misinformation.  

Lack of Context in Edge Cases: The model might struggle with problems that rely on highly specialized domain knowledge or cultural-specific contexts not covered in training.  

Privacy and Security Concerns: Applications involving sensitive data could pose risks if not managed with strict data security protocols, raising concerns about data misuse or breaches.  

#2 – NVIDIA Became Most Valuable AI Company

In November, NVIDIA became the most valuable company in the world, surpassing Apple and highlighting the market demand for AI hardware.

At its peak, NVIDIA’s market capitalization surpassed $3.3 trillion, and its stock price nearly tripled in value throughout 2024 as investors grew more confident in its ability to maintain its growth rate amid the AI boom.

Over the last five years, its stock grew by over 2,700% and its revenue has more than doubled in each of the last five quarters—tripling in the last three—as it designs about 75% of the world's AI chips.

It did briefly push past Apple in June (but just for a day), and it also beat Microsoft (valued at $3.06T) in October, as it continued its “astronomical” upward trajectory.

This just further highlights the increasing importance and dominance of AI infrastructure within the industry, something which is expected to continue, although whether NVIDIA will remain the largest AI chip manufacturer and, therefore, the most valuable company in the world, remains to be seen, with many (such as OpenAI, Apple, Meta, and Microsoft) rumored to be developing their own AI chips, to reduce reliance on NVIDIA.

#3 – The EU launches the AI Act

In March, the European Union (EU) announced it was developing new risk-based legislation—The AI Act—to regulate the development, use, and applications of AI in the EU, which it passed in May

The EU AI Act is a landmark piece of legislation that aims to establish a comprehensive regulatory framework for artificial intelligence (AI) within the European Union. It came into force on August 1, 2024. 

The EU gave tech companies 3-6 months to comply with the new rules, or they will face potential fines ranging from $8.1M (or 1% of their global annual turnover) to $38M (or 7% of their global annual turnover).

Key Provisions

• Risk-Based Approach: The Act categorizes AI systems into different risk levels.

o Unacceptable Risk: AI systems deemed to pose an unacceptable level of risk, such as those that manipulate people into harmful choices or exploit vulnerabilities, are prohibited.

o High-Risk: High-risk AI systems, such as those used in critical infrastructure, law enforcement, or education, are subject to strict requirements, including conformity assessments and risk management systems.

o Limited Risk: AI systems with limited risk, such as spam filters or video game AI, are subject to minimal regulation.

o Minimal Risk: AI systems with minimal risk, such as AI-powered chatbots or spam filters, are largely unregulated.

• General-Purpose AI Models: The Act introduces specific provisions for general-purpose AI (GPAI) models, such as large language models, requiring providers to implement risk mitigation measures and ensure the models' compliance with the Act's requirements.

• Transparency and Accountability: The Act emphasizes the importance of transparency and accountability in the development and deployment of AI systems. It requires providers to provide users with clear information about the AI systems they interact with and to be accountable for the systems' compliance with the Act's provisions.

• Enforcement: The Act empowers national authorities to enforce its provisions and impose significant penalties for non-compliance.

The EU AI Act is expected to have a significant impact on the development and deployment of AI systems globally. It sets a high bar for AI safety, ethics, and transparency, potentially influencing the development of AI regulations in other jurisdictions. 

Concluded.

Disclaimers: Pictures in these blogs are taken from free resources at Pexels, Pixabay, Unsplash, and Google. Credit is given where available. If a copyright claim is lodged, we shall remove the picture with appropriate regrets.

For most blogs, I research from several sources which are open to public. Their links are mentioned under references. There is no intent to infringe upon anyone’s copyrights. If, however, it happens unintentionally, I offer my sincere regrets.

Reference:

https://aitoolreport.beehiiv.com/