Tether CEO Issues Scary AI Warning

As a crypto investor with experience in the tech industry, I share Paolo Ardoino’s concerns about centralized large language models (LLMs) like OpenAI. The security breach at OpenAI earlier this year was indeed scary, and it raises serious questions about the adequacy of their security measures. The fact that they chose not to disclose the breach until it was leaked to the media only adds to the alarm.


Paolo Ardoino, the CEO of Tether, cautioned on X social media platform about the potential risks associated with centralized large language models.

Ardoino expressed concern over the significant security incident that occurred at OpenAI, the prominent generative AI company, in early 2023, labeling it as “alarming.”

A recent report by The New York Times revealed that OpenAI withheld information about a data breach wherein some sensitive details were leaked.

Previously employed researchers at OpenAI, including Leopold Aschenbrenner, have raised concerns about the company’s insufficient security protocols, potentially leaving it susceptible to malicious intrusions from foreign entities. Aschenbrenner asserted that his dismissal was due to political differences. However, OpenAI disputed this claim, stating that the security breach was disclosed prior to Aschenbrenner joining their team.

Despite OpenAI’s assertion that there are no national security concerns with their current technology, some remain apprehensive about the potential risk of its secrets being accessed by China.

As an analyst, I’ve noticed that besides security breaches, there are valid criticisms raised against centralized AI models regarding unethical data usage and censorship. The CEO of Tether advocates for the utilization of local AI models as the most effective solution to tackle privacy concerns and ensure both resilience and independence.

In a recent post on the X social media platform, Ardoino expressed that locally running artificial intelligence (AI) models is the most effective method for safeguarding individuals’ privacy and achieving both resistance and self-sufficiency.

He has added that modern smartphones and laptops are powerful enough for fine-tuning general LLMs. 

Read More

2024-07-08 09:35