Good morning, it's Friday. AI is getting louder, faster, hungrier—like a model trained on Daft Punk. ElevenLabs just secured $180M to make synthetic voices indistinguishable from human ones, NVIDIA's RTX 5090 is shattering benchmarks, and California is flirting with nuclear power to feed AI's insatiable energy appetite.
Meanwhile, Google quietly dropped Gemini 2.0 Pro Experimental, and if you thought AI art could be copyrighted, the US Copyright Office just said, "Nice try."
Read on!
🤔 FRIDAY FACTS
What was the first AI to beat a world champion in a game of chess?
Stick around to find out the answer! 👇
☝️ POWERED BY RECRAFT V3
Recraft V3 Is SOTA in Image Generation
Recraft V3 sets a new standard in image generation, outperforming all competitors according to Hugging Face’s Text-to-Image Benchmark by Artificial Analysis. The main advantages of Recraft V3 are text generation quality, anatomical accuracy, prompt understanding, and aesthetic quality. Recraft V3 is the only model in the world that generates images with long texts. Now available via API with unique functionalities no other tool provides: the ability to generate image sets and vector art.
Use BERMAN12 for a $12 discount on any paid plan—start creating with Recraft now!
🦖 TECH MONSTER
Michael Crichton’s Warning: When Technology Becomes the Monster
The Recap: Michael Crichton’s approach to storytelling—where technology, rather than individuals, drives the plot—offers a crucial lens for understanding Big Tech and artificial intelligence today. Cal Newport argues that our focus on tech billionaires like Elon Musk and Mark Zuckerberg misses the larger issue: technologies often escape human control, creating unintended consequences far beyond their creators' intentions.
Highlights:
Crichton’s editor Robert Gottlieb advised him to strip away character depth in The Andromeda Strain, making technology itself the protagonist—an approach that shaped his later works like Jurassic Park.
Unlike Mary Shelley’s Frankenstein, which emphasizes personal tragedy, Jurassic Park is about how technology evolves beyond its makers’ grasp, making it a superior cautionary tale.
In Jurassic Park, John Hammond isn’t a sinister mastermind but a naive entrepreneur who fails to foresee the chaos he unleashes, mirroring Silicon Valley’s reckless innovation.
The obsession with personalities—Musk’s erratic behavior, Zuckerberg’s “masculine energy”—distracts from deeper structural issues with social media and AI.
Twitter/X was designed for mass engagement, but its core mechanics naturally promote outrage and misinformation—problems that persist regardless of ownership.
Like social media, email became overwhelming not because of bad bosses but because its frictionless design encouraged endless communication. AI could follow the same path, becoming unmanageable despite good intentions.
Instead of focusing on tech moguls, we should ask whether these technologies serve us, and if not, how we can “blow them up” metaphorically through regulation, alternative designs, or outright abandonment.
Forward Future Takeaways:
Crichton’s stories highlight the power and unpredictability of emerging technologies, but rather than fearing AI as the next uncontrollable force, we should embrace its potential. Unlike Jurassic Park’s dinosaurs, AI isn’t a rogue entity—it’s a tool we actively shape. The real lesson isn’t about halting progress but ensuring that AI serves humanity’s best interests. Instead of waiting for crises to dictate change, we should proactively build AI systems that are transparent, ethical, and aligned with societal goals. → Read the full article here.
⚠️ CYBER THREATS
How Threat Actors Are Testing the Limits of Generative Models
The Recap: Google’s Threat Intelligence Group (GTIG) has released a comprehensive analysis of how government-backed cyber threat actors and information operations (IO) actors attempted to misuse Google’s AI-powered assistant, Gemini. While adversaries are experimenting with AI for research, content generation, and coding assistance, the report finds no evidence that AI has fundamentally changed cyber threats—at least not yet.
Highlights:
Threat actors are using Gemini for common cyber activities like reconnaissance, research, and scripting, but they have not developed novel AI-driven attack techniques.
Iranian APT actors were the most active users, leveraging Gemini for reconnaissance, phishing campaigns, vulnerability research, and content manipulation for influence operations.
Chinese APT actors used Gemini to research U.S. military networks, reverse engineer security tools, and troubleshoot scripts for privilege escalation and data exfiltration.
North Korean actors used AI to research jobs, draft cover letters, and generate freelance proposals, likely supporting efforts to place IT workers in Western companies under false identities.
Unlike other nations, Russian APT actors showed minimal use of Gemini, possibly due to security concerns or preference for domestically controlled AI models.
Threat actors attempted basic jailbreaks using publicly available prompts but failed to bypass Gemini’s safety controls. AI proved useful for productivity gains but did not enable breakthrough capabilities.
Malicious actors are promoting jailbroken LLMs such as "FraudGPT" and "WormGPT" for phishing and malware development, indicating a growing black market for AI-assisted cybercrime.
Forward Future Takeaways:
While AI is accelerating cyber operations, it has not yet revolutionized them—most threat actors still rely on traditional hacking methods with AI as an efficiency booster. However, as AI models grow more advanced and new agentic systems emerge, the threat landscape is expected to evolve. Google emphasizes the need for robust AI security frameworks, such as its Secure AI Framework (SAIF), to mitigate future risks. The takeaway? AI isn't yet a cyber "superweapon," but proactive defense is crucial to ensure it stays that way. → Read the full article here.
🤐 AI CENSORSHIP
DeepSeek’s AI Chatbot Struggles with Censorship, In Real-Time
The Recap: DeepSeek, the Chinese AI chatbot, provides glimpses of censored information before abruptly deleting its own responses, revealing the delicate balancing act it performs under China’s strict internet controls. While occasionally offering insights that are more open than domestic social media, the chatbot ultimately adheres to government narratives, self-censoring in both English and Chinese—with the latter being more restrictive.
Highlights:
When asked about China’s “Zero Covid” policies, DeepSeek initially provided a nuanced response, mentioning protests and public dissatisfaction—before wiping its own answer and replacing it with a vague disclaimer.
Questions about Xi Jinping or other high-ranking Chinese officials resulted in automatic refusals, with no biographical details provided, while lower-ranking figures received partial responses—sometimes deleted mid-answer.
Inquiries about the war in Ukraine led to two different approaches: in English, the chatbot acknowledged Russia’s "full-scale invasion", but in Chinese, it softened the language to align with China’s official stance.
The chatbot exhibited real-time self-censorship, modifying answers to be progressively less sensitive when asked the same question repeatedly.
Certain censorship workarounds—like using letter substitutions to reference the Tiananmen Square Massacre—were possible initially but were quickly patched by DeepSeek’s developers.
On the topic of China’s internet censorship, DeepSeek openly acknowledged state regulations but framed them as necessary for “cybersecurity and social stability,” carefully avoiding the term “censorship.”
Despite China’s Great Firewall usually limiting censorship to domestic users, DeepSeek appears to enforce the same content restrictions internationally—potentially exporting Beijing’s digital control methods abroad.
Forward Future Takeaways:
DeepSeek’s self-censorship highlights the tension between AI's capacity for open-ended reasoning and China’s strict control over information. The bot’s behavior suggests an evolving struggle, with developers actively closing loopholes as they arise. If DeepSeek gains global traction, it could spread China’s controlled narratives beyond its borders, reinforcing the growing intersection of AI and state propaganda. → Read the full article here.
🛰️ NEWS
Looking Forward
🎵 Riffusion’s AI Music Tool Is Free: The startup’s new AI, Fuzz, creates personalized songs from text, audio, or images. Unlike competitors, Riffusion offers its platform completely free worldwide.
⚛️ OpenAI’s AI to Aid Nuclear Research: OpenAI will provide its models to U.S. National Labs for nuclear security and scientific projects. The partnership, involving Microsoft, includes deployment at Los Alamos.
🔓 DeepSeek AI Exposed Sensitive Data: The rising Chinese AI startup left a database open, leaking secret keys, chat logs, and backend details. Security experts warn of major risks in AI data protection.
📽️ VIDEO
Forward Future’s First CES!
Matt shares his first CES experience, spotlighting NVIDIA’s RTX 5090, AI-powered gaming, and cutting-edge autonomous technology. He explores NVIDIA’s open-source projects, advanced robotics, and AI-driven video editing tools. Get the full scoop in Matt’s latest video! 👇
🧰 TOOLBOX
Effortless Design, Realistic Avatars, and Precision Video Editing
Adobe Express | AI-Powered Design: Adobe Express simplifies content creation with templates, AI tools, and one-click editing for individuals and teams.
Humva | AI Avatars for Videos: Humva creates hyper-realistic AI presenters with customizable appearance, voice, and gestures for professional video content.
Go-with-the-Flow | AI Video Precision: Go-with-the-Flow enhances video editing with AI-driven animation control, seamless transitions, and advanced effects.
🤔 FRIDAY FACTS
Deep Blue Became the First AI To Defeat a Reigning World Chess Champion
The first AI to defeat a reigning world chess champion was IBM’s Deep Blue, which beat Garry Kasparov on May 11, 1997.
Deep Blue was a specialized supercomputer capable of evaluating 200 million positions per second—a level of brute-force calculation no human could match. Kasparov, widely regarded as one of the greatest chess players in history, had previously defeated Deep Blue in 1996. But after an upgrade, the machine came back stronger and won the rematch 3.5–2.5.
The match was a turning point in AI history, marking the first time a machine proved it could outplay the best human in a game once thought to require intuition and creativity. Today’s AI chess engines, like Stockfish and AlphaZero, make Deep Blue look primitive—but in 1997, it was the checkmate heard 'round the world. ♟️
📥 FF INTEL
Got a Hot Tip or Burning Question?
We’re all ears. Drop us a note, and we’ll feature the best reader insights, questions, and scoops in future editions. Let’s build this thing together.
🍵 Hit the button below and spill the tea!
🤝 CONNECT
Stay in the Know
Follow us on X for quick daily updates and bite-sized content.
Subscribe to our YouTube channel for in-depth technical analysis.
Prefer using an RSS feed? Add Forward Future to your feed here.
Thanks for reading today’s newsletter. See you next time!
The Forward Future Team
🧑🚀 🧑🚀 🧑🚀 🧑🚀
