Ilya Sutskever breaks silence on AI’s future

Good morning, AI enthusiasts. The AI industry is betting tons of of billions on one idea: more compute equals smarter AI.

However the man who helped construct ChatGPT says that era is ending. OpenAI co-founder Ilya Sutskever just broke his silence in a rare interview — sharing his tackle ASI timelines, his secretive startup, and why research will drive the following AI leap.

In today’s AI rundown:

• Ilya Sutskever says AI’s ‘age of scaling’ is ending

• Black Forest Labs’ Flux.2 image generation suite

• Use ChatGPT shopping research to seek out best deals

• Anthropic: AI could double U.S. productivity growth

• 4 latest AI tools, community workflows, and more

LATEST DEVELOPMENTS

ILYA SUTSKEVER

🧠 Ilya Sutskever says AI’s ‘age of scaling’ is ending

The Rundown: Secure Superintelligence founder Ilya Sutskever just appeared on the Dwarkesh Podcast, giving his tackle scaling, ASI, his secretive startup, and more — arguing that research breakthroughs, not compute, will drive the following wave of progress.

The small print:

• Sutskever said that 2020-2025 was the “age of scaling”, but we’ve reached the purpose where research becomes the differentiating factor for AI breakthroughs.

• He forecasts 5-20 years until superhuman-like learning AI emerges, adding that the primary ASI systems needs to be built to care about sentient life.

• Sutskever said that his startup, SSI, is taking a “different technical approach” to superintelligence, and called it an “age of research” company.

• He also revealed that SSI was raising at a $32B valuation and declined an acquisition offer from Meta, along with his cofounder marking the one departure.

Why it matters: Sutskever has been out of the highlight since his exit from OpenAI, with SSI quietly working within the shadows — but his words carry massive weight within the AI world. His tackle a “return to research” over compute comes at a clumsy time, as the vast majority of the industry continues to pour massive money into scaling infrastructure.

TOGETHER WITH GURU

🧠 Your AI source of truth

The Rundown: Guru is the AI Source of Truth that connects your whole company’s tools and delivers cited, permission-aware answers in every single place you’re employed. With one governed knowledge layer powering each your people and your AIs, teams move faster — with fewer blind spots and mistakes.

Guru means that you can:

• Connect all knowledge with permission-aware access

• Get trusted, cited answers in chat and in every single place else you’re employed

• Experience knowledge that improves and verifies itself

Learn more about Guru.

BLACK FOREST LABS

🎨 Black Forest Labs’ Flux.2 image generation suite

The Rundown: Black Forest Labs dropped Flux.2, a brand new family of powerful image models — featuring multi-reference capabilities that maintain character and magnificence consistency across as much as ten input images and value reductions in comparison with rivals.

The small print:

• FLUX.2 combines a model that handles each text and pictures with one other that handles spatial relationships for realistic lighting, physics, and compositions.

• The models are available barely below Google’s recently released SOTA Nano Banana Pro, but offer a big cost reduction in pricing.

• The lineup includes Pro for top-quality API access, Flex for dev customization, Dev as an open-weights option, and Klein coming soon as fully open-source.

• Outputs now reach as much as 4MP with improved typography capabilities, enabling production-ready infographics, UI mockups, and sophisticated text layouts.

Why it matters: Nano Banana Pro felt like a step change within the range of creative workflows and talents, but Flux.2 shows the competition isn’t lagging far behind. While AI’s image realism was already virtually imperceptible from reality, the next-gen world knowledge, consistency, and text capabilities are the following step forward.

AI TRAINING

🏷️ Use ChatGPT shopping research to seek out best deals

The Rundown: On this tutorial, you’ll learn find out how to use ChatGPT’s Shopping Research feature to seek out one of the best deals without hopping across multiple sites, centralizing your entire experience from search to cost comparison to checkout links.

Step-by-step:

1. Go to ChatGPT, click the plus button, and choose Shopping Research to enable the tool

2. Describe what you are on the lookout for (e.g., “Find me one of the best Black Friday laptop deals and compare across sites”)—answer follow-up questions on budget, specs, and use case to refine results

3. Review final suggestions showing top deals, price comparisons across retailers, trade-offs, alternatives, and direct links to every listing

4. Select your product, click through to the retailer, and complete your purchase

Pro Tip: Soon, you’ll also give you the chance to purchase chosen products directly in ChatGPT via services like Stripe or PayPal — no must visit the retailer’s site.

PRESENTED BY INCOGNI

📞 Unknown number calling?

The Rundown: Scammers don’t pick phone numbers at random – they buy your data from brokers. The BBC caught call center employees on hidden camera, with one bragging about making $250K from victims.

With Incogni, you possibly can:

• Robotically delete your personal information from data brokers’ databases

• Monitor and block latest data exploits as they seem across the net

• Reduce spam calls, phishing attempts, and identity theft risk repeatedly

Try Incogni here and get 55% off your subscription with code RUNDOWN.

AI RESEARCH

📈 Anthropic: AI could double U.S. productivity growth

The Rundown: Anthropic published latest research analyzing 100K Claude conversations to trace AI’s productivity gains, estimating that widespread AI adoption could boost annual U.S. labor productivity growth by 1.8% — doubling the present rate.

The small print:

• Anthropic researchers fed 100K anonymized conversations through its Clio privacy tool, mapping tasks to federal labor data to calculate productivity gains.

• Researchers found Claude cuts task completion time by roughly 80%, with the common work request taking about 90 minutes without assistance.

• Software developers account for 19% of estimated productivity gains, followed by operations managers, marketing specialists, and customer support roles.

• Examples of tasks with massive time savings included curriculum development (96%), research assistance (91%), and executive admin functions (87%).

Why it matters: There may be loads of debate over AI’s actual impact vs. hype, and this research shows the true gains across a wide range of sectors and tasks. But the larger query the study sidesteps: whether the estimated doubling of productivity growth comes with the job displacement Anthropic’s own CEO continues to warn about.

QUICK HITS

🛠️ Trending AI Tools

• 🤖 Claude Opus 4.5 – Anthropic’s latest benchmark-topping frontier model

• 🎆 FLUX.2 – Black Forest Labs’ latest visual intelligence model

• 🛒 Shopping Research – Latest interactive shopping experience in ChatGPT

• 👁️ HunyuanOCR – Tencent’s open-source, SOTA visual understanding model

📰 All the things else in AI today

Nvidia responded to concerns over Google’s TPUs gaining a foothold, saying its hardware is “a generation ahead” with “greater performance, versatility, and fungibility.”

Anthropic tested Claude Opus 4.5 on a take-home exam given to prospective performance engineers, with the AI scoring “higher than any human candidate ever.”

AI music platform Suno partnered with Warner Music Group to coach on licensed recordings and let users create songs with participating artists’ voices and styles.

Google’s Gemini 3 Pro set a brand new high rating for AI models with a 130 on Tracking AI’s offline IQ test, surpassing Grok 4 Expert Mode’s 126.

Tencent’s Hunyuan open-sourced HunyuanOCR, a SOTA visual understanding model for document parsing, information extraction, text detection, and more.

Perplexity launched a free AI shopping feature for U.S. users that learns personal preferences and enables purchases directly throughout the app through PayPal.

COMMUNITY

🤝 Community AI workflows

Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.

Today’s workflow comes from reader Tara R. within the U.K.:

“I exploit NotebookLM to create podcasts for my students, as a method to refresh an entire term’s price of teaching. It’s a practical way for them to review topics while they’re on the go, especially great for individuals who struggle to seek out time to revise.”

How do you utilize AI? Tell us here.

🎓 Highlights: News, Guides & Events

• Read our last AI newsletter: Anthropic enters the frontier AI fight

• Read our last Tech newsletter: This device wiretaps your ‘second brain’

• Read our last Robotics newsletter: Figure sued over ‘skull-crushing’ force

• Today’s AI tool guide: Use ChatGPT shopping research to seek out deals

• Watch our last live workshop: The Human-First Agentic Content Workflow

See you soon,

Rowan, Joey, Zach, Shubham, and Jennifer — the humans behind The Rundown