Monday’s top story: AI startup—Hugging Face—has been targetted by hackers, comprising several of their user’s sensitive and private information.
🔒 Hugging Face hackers
🔍 NVIDIA's 2026 AI breakthrough
🔊 Sound effects with ElevenLabs?
📽️ Sora premieres at film festival
Read Time: 5 minutes
Our Report: Hugging Face—an AI startup with a massive repository of AI apps created and submitted by its community—detected suspicious, “unauthorized access” to Spaces, its platform for creating, sharing, and hosting AI models and resources.
🔑 Key Points:
Hackers gained access to users' authentication ‘secrets’, which hold sensitive data like the tokens and credentials used to log in and communicate with 3rd party apps, securely.
Hugging Face has revoked compromised authentication tokens, emailed those affected, and recommended that Space users switch to ‘fine-grained access’ tokens, with tighter controls
They’re also working with “cyber security forensic specialists” to investigate the incident and have already reported it to law enforcement and data protection agencies.
🤔 Why you should care: Although it’s not clear how many users or apps were impacted by this breach, it comes as Hugging Face faces increasing scrutiny over its security practices, with cybersecurity firms highlighting vulnerabilities and apps, within the platform, capable of adding malicious code onto users’ machines.
Together with Patent
Our bi-weekly newsletter uncovers where tech titans (like Apple, Meta, and Tesla) are investing, by scouring the U.S. Patent & Trademark website.
Get proprietary news and analysis of innovation at its earliest stage.
Guidde is the secret presentation tool that will 10x your team’s productivity 🤫💡
Otocom is a personalized eCommerce recommendations chatbot
FoxyApps creates AI-powered applications with ChatGPT
IX Coach is an AI-powered growth development coach
Aithenticate uses AI to boost website credibility
Our Report: During a speech at the Computex trade show in Taipei, NVIDIA CEO, Jensen Huang, revealed its Blackwell AI chip successor (announced just a few months ago): 'Vera Rubin', which is expected to launch in 2026.
🔑 Key Points:
Although details are light, Rubin will have new graphics processors (GPUs) and central processors (CPUs) designed to power AI applications with high-bandwidth memory.
Alongside the introduction of Rubin, Huang also announced that NVIDIA will upgrade its AI accelerators every year, revealing that they’re launching the Blackwell Ultra chip launch in 2025.
Huang said NVIDIA’s chips will produce 98% cost savings and use 97% less energy than “traditional computing methods” that can’t process the data needed to power AI systems.
🤔 Why you should care: Announcing a new next-generation chip and accelerating its release schedule from every two years to every year, it's clear that NVIDIA is making a bold move to secure its dominance over the AI chip market, which it currently has an 80% share of.
Use the Prompt Engineer GPT by AI Tool report to 10x your prompts.
✅ Facebook ads
copy & paste ⬇️
“I'm looking for a Facebook ad copy that will use the social proof and credibility of [influencer type] to persuade my [ideal customer persona] to try my [product/service] and share their positive experience with their followers.”
Our Report: After announcing the project back in February, voice cloning startup–ElevenLabs–has launched a new AI tool that allows users to use prompts to generate sound effects.
🔑 Key Points:
Users can type prompts like “waves crashing,” or “metal clanging,” to generate snippets of sounds and can also generate musical clips (max 22 secs) with prompts like “guitar loops.”
Users who don’t have a subscription will still be able to generate around 60 sound effects p/m but will have to add “elevenlabs.io” if they publish content with the sound effects.
Shutterstock’s audio library—containing licensed tracks—was used to train the tool, and it’s been tested by video game developers, film producers, and social media content creators.
🤔 Why you should care: ElevenLabs has made it clear that the tool will not allow any prompts that violate its Prohibited Content and Uses Policy, which includes things like self-harm, child safety threats, and fraud.
Our Report: The Tribeca Film Festival (based in New York and founded by Robert de Niro) is dedicating a whole section to films created by AI, featuring “Sora Shorts”, generated by OpenAI’s text-to-video platform, Sora.
🔑 Key Points:
The film festival will premiere five short films, generated by Sora, on June 15th, featuring films produced by the red teamers with Sora access, including “Nanny” director Nikyatu Jusu.
The five filmmakers were allowed to create their videos independently but were given just a few weeks to complete their short films, ahead of the festival in June.
OpenAI’s COO—Brad Lightcap—said that they’re keen to see the films and learn how they can make Sora better for creatives and “are honored for their works to debut at the Tribeca Festival.”
🤔 Why you should care: While we don't know how long these Sora Shorts will be, Sora can currently produce just 60 seconds of film at a time, so it’s thought that the five filmmakers will stitch clips together to make a lengthier film, so it’ll be interesting to see if Sora is capable at replicating characters and scenes from previous prompts.
Meta is feeding its user comments into its AI system to generate AI summaries, which is causing some privacy concerns.
Voice researchers have confirmed that OpenAI’s AI voice option—Sky—does sound like Scarlett Johannson, and also contains elements of Anne Hathaway and Keri Russell’s voices, too.
Google Meet researchers are working on “ChatDirector,” which will rival Microsoft Teams, and allow users to use 3D aviators.
Hit reply and tell us what you want more of!
Until next time, Martin & Liam.
P.S. Don’t forget, you can unsubscribe if you don’t want us to land in your inbox anymore.