March 17, 2023

Nvidia is getting it right. We’re getting it wrong. Bots are getting away.

hardware

Nvidia: Intrinsic Valuation Is Spicy With AI Hype

Seeking Alpha, 03/17/23. The popularity of Nvidia’s high-performance GPUs have driven the company’s stock price by 111.4%. These products are considered the backbone of the AI industry. 

Nvidia’s A100 and H100 GPUs are vital for AI supercomputers with the latter offering significant performance improvements. The company has partnered with major players such as Microsoft, Meta Platforms, and Oracle. While it’s revenue from gaming has declined, there are signs of recovery. 

An intrinsic valuation model shows the stock as 24% overvalued. It’s currently labeled as “hold” as it approaches a technical resistance level. Investors may consider buying back if the stock price falls below intrinsic valuation. READ MORE

ai

One Thing We’re Getting Wrong About AI

CNET, 03/17/23. AI’s rapidly increasing popularity is generating fascination and concern. Its capabilities and limits may not be fully understood. And sensationalizing AI may create fear and undermine trust. 

With big tech companies joining the AI race, tools like OpenAI’s ChatGPT make AI accessible to average users. However, people incorrectly attribute sentience to AI. It is also reflecting biases and generating incorrect information.

Ethical safeguards and human rights principles must be embedded into AI development. Despite its potential, AI-generated content may lack the human touch of human-created art.

READ MORE

IDENTITY

AI Bots May Have Defeated CAPTCHA Tests for Good

INVERSE, 03/17/23. Captcha is a tool to distinguish humans from machines for cybersecurity purposes, developed in the 1980s to trick internet forum bots and in 1997 to prevent data theft. 

Captcha uses distorted text and numbers to challenge bots, but hackers can bypass them using solver services. Future systems may use AI or other technologies, and an alternative called Turnstile checks browser behavior in a second. READ MORE

GPT-4

Will GPT-4 change the world?

THE WEEK, 03/17/23. OpenAI has released its latest large language model, GPT-4, which can handle more words, generate fewer errors, and is less likely to produce disallowed content. 

GPT-4 is multi-modal, capable of analyzing both images and text. It has the potential to enhance services such as language learning and customer service. However, the technology still can generate biased, false, and hateful text. Further questions revolve around its potential impact on data privacy and power consolidation. READ MORE

MICROSOFT

Microsoft is adding an A.I. Copilot to its office software that can summarize long emails and draft stories

FORTUNE, 03/17/23. Microsoft is bundling AI tools into its suite of office software including Word, Excel, and Outlook emails. The new feature, called Copilot, will help users summarize long emails, draft stories, and animate slides. 

Microsoft is positioning Copilot as a tool that will make workers more productive by freeing up time spent on emails and analysis. Additionally, the company intends to add a chat function to leverage user data to carry out actions. Google has already added generative AI tools into its Workspace applications. READ MORE