Instagram story viewer> @thenexttoken> Posts

The Next Token

@thenexttoken

11

followers

5

following

Daily AI Podcast

POSTS STORIES REELS TAGGED

#ai #artificalintelligence #voiceagent #openclaw by @thenexttoken

3

0

8 months ago

For the latest episode of ‘The Next Token’ (available on Spotify & Apple Podcasts), I’m breaking down the most important battle going on in tech today: the AI Chip War.

The race for AI dominance has become a hardware race. This has sparked a geopolitical battle between the US and China over the high-performance GPUs needed to build the future.

In the episode, I walk through how the US is using export controls and key alliances (like with ASML in the Netherlands) to create a ‘two-tiered’ tech world, keeping cutting-edge chips from China and how a massive black market for smuggled GPUs and an expensive national mission to achieve semiconductor independence have become a high priority for China.

This raises a critical question for the future of AI innovation: Will these controls maintain a Western lead, or just create a powerful, independent rival with a grudge?

#ai #geopolitics #Tech #GPU #NVIDIA #Podcast #TheNextToken by @thenexttoken

4

0

a year ago

For the first episode of my new podcast ‘The Next Token’ (available on Spotify/apple podcasts) I spent a few flux/runway credits on making a short video snippet.

The first episode covers the latest threat intelligence report from Anthropic that confirms that AI is already being weaponized for sophisticated cybercrime.

Through using clever prompt engineering (pretending the AI user is doing official security testing or white-hat hacking among other techniques), the ethical guardrails of the Claude models can be jailbroken. This is concerning as Anthropic are one of the best at AI-safety. If this can happen with closed-source air-tight thoroughly tested models, the danger of open-source models (which have comparable ability but very easily jail-broken guardrails) is extreme.

In the episode I walk through three different ways AI is being used in cybercrime as detailed in the report

The Force Multiplier: A single hacker can now operate with the impact of an entire criminal syndicate (vibe hacking), using AI to autonomously run attacks from reconnaissance to extortion.

Simulated Competence: State-sponsored operatives are using AI to pass interviews and hold down high-paying remote tech jobs they aren’t qualified for, turning sanctions evasion into a scalable enterprise for the North Korean Regime.

The Criminal Marketplace: The barrier to entry for creating and selling advanced Ransomware-as-a-Service has effectively collapsed, making sophisticated tools available to low-skill actors.

This raises a critical question for every tech company: Are your hiring processes and security protocols prepared for an era of AI-simulated competence?

Slightly edited extract from the end of the podcast:

‘By transparently sharing these cases of misuse, Anthropic is sounding the alarm bells for the entire industry.

This isn’t just a list of hacks, it’s an urgent warning shot from the front line in the AI safety war from one of the leaders.

It’s an important reminder that the race to build powerful AI must also be a race to secure it.’

#AI #Cybersecurity #Anthropic #Artificialintelligence #Podcast #TheNextToken by @thenexttoken

7

0

a year ago