Greetings Qrew,

Whether you're diving into AI for work, side projects, or just out of curiosity — this issue of The Creators Hive brings you practical insights, hands-on learning, and tools you can actually use.

1. AI ‘godmother’ advocates for spatial intelligence
2. Anthropic’s big cost advantage over OpenAI
3. GPT-5 cracks a full 9x9 Sudoku puzzle
4. OpenAI’s reccos to brace for superintelligent AI
5. McKinsey’s 2025 AI reality check
6. Tool Recommendation of the Day: Jigidi

---

AI ‘godmother’ advocates for spatial intelligence

Dr. Fei-Fei Li, often called the “godmother of AI,” published a new essay arguing that the next major leap in artificial intelligence will come from spatial intelligence — the ability for systems to understand, reason about, and generate physics-consistent 3D environments.

Key details:

• Li noted that while language models excel at abstract reasoning, they lack spatial perception — such as estimating distance, motion, or object relationships.
• She described spatial understanding as the cognitive foundation of human intelligence and essential for bridging the gap between thinking and acting.
• Li emphasized that world models capable of generating and interpreting realistic 3D spaces, visual inputs, and cause-effect relationships are key to this progress.
• Such models could power breakthroughs across robotics, healthcare, science, and design, enabling AI to reason and interact meaningfully with the physical world.

Why it matters: Spatial intelligence may define AI’s next evolutionary phase — moving beyond words and data toward embodied understanding of the real world. Systems that grasp physics and 3D interactions could model chemical reactions, predict climate dynamics, and even simulate new technologies before they’re built.

---

Anthropic’s big cost advantage over OpenAI

Anthropic reportedly forecasts a massive cost advantage over OpenAI — projecting significantly lower compute spending to train and operate its AI models over the next several years, according to The Information.

The details:

• Anthropic expects to spend $6B on compute in 2025 versus OpenAI’s $15B, with totals rising to $27B by 2028 compared to OpenAI’s $111B.
• The company achieves savings by using chips from Amazon, Nvidia, and Google for different workloads, unlike OpenAI’s heavier reliance on Nvidia GPUs.
• Anthropic also raised its revenue projections, targeting $70B in revenue by 2028 and cash flow positivity by 2027.
• OpenAI, by contrast, aims for $100B in 2028 revenue but isn’t expected to turn cash flow positive before 2030.

Why it matters: Anthropic’s focus on efficiency and enterprise partnerships (with 80% of revenue coming from API clients) gives it a more sustainable path — while OpenAI’s broader consumer and product-driven approach demands greater spending. The contrast highlights two distinct philosophies shaping the AI race: scale versus sustainability.

---

GPT-5 cracks a full 9x9 Sudoku puzzle

GPT-5 has become the first AI model to solve a complete 9x9 Sudoku puzzle, according to Sakana AI’s Sudoku-Bench — a benchmark built to test deep reasoning, spatial logic, and creativity.

Key details:

• Sudoku-Bench, launched in May, challenges large language models with both classic and hybrid Sudoku variants that require multi-step logical reasoning.
• Until now, no AI had fully solved a 9x9 Sudoku puzzle — GPT-5 is the first to do so, demonstrating major gains in structured reasoning.
• GPT-5 recorded a 33% solve rate across puzzles, nearly double that of the previous top-performing model.
• However, 67% of puzzles remain unsolved, showing continued difficulty with “meta-reasoning” — the ability to adapt to new rule sets and creative problem-solving patterns.

Why it matters: GPT-5’s Sudoku success highlights progress in systematic and spatial reasoning — a crucial step toward more human-like intelligence. Still, the gap between logical precision and creative intuition remains wide, marking the next frontier in AI’s cognitive development.

---

OpenAI’s reccos to brace for super intelligent AI

OpenAI has outlined its vision for managing the rise of superintelligent AI, predicting that AI systems will soon be capable of making scientific discoveries and urging global coordination on safety, regulation, and resilience.

Key details:

• OpenAI stated that current AI systems already outperform top humans in complex intellectual tasks and are “about 80% of the way to an AI researcher.”​
• The company expects AI to make minor scientific discoveries by 2026 and major breakthroughs by 2028, as intelligence costs decline by roughly 40x annually.
• For the coming era of superintelligence, OpenAI emphasized the need to collaborate with governments and safety agencies to mitigate risks such as bioterrorism or uncontrolled self-improvement.
• It also called for shared safety standards among leading labs, a resilience ecosystem akin to cybersecurity, and continuous monitoring of AI’s real-world effects to guide policymaking.

Why it matters: OpenAI’s roadmap signals that superintelligent AI may arrive sooner than many expect. Preparing global safety protocols, oversight mechanisms, and resilience systems could determine how humanity manages — or is managed by — the next stage of artificial intelligence.

---

McKinsey’s 2025 AI reality check

​McKinsey’s State of AI 2025 report, surveying nearly 2,000 organizations, reveals that while AI adoption is nearly universal, most companies remain stuck in pilot phases, with few realizing enterprise-wide transformation or scaling intelligent agents.

The details:

• 88% of companies now use AI in some form, but only 33% have scaled it beyond pilot projects.
• 39% of firms reported EBIT gains from AI, yet only 6% achieved impacts exceeding 5%, primarily by reengineering workflows and driving innovation.
• 62% are exploring AI agents — though just 23% have scaled them, mostly in IT and knowledge management.
• 32% of companies anticipate workforce reductions of 3% or more next year, while 13% expect to grow headcount, with larger firms more likely to forecast cuts.

Why it matters: The report underscores a widening execution gap. The few companies achieving measurable returns aren’t merely automating — they’re redesigning operations, integrating AI across functions, and using it to create new value rather than just trim costs.

---

Tool Recommendation of the Day: Jigidi

Website:jigidi.com

What makes it unique

Jigidi is a free online jigsaw puzzle community where users can both solve puzzles and create their own from images. It blends puzzle-solving with social interaction, allowing you to share, comment on, and explore puzzles made by others.

Key Features

• A large catalog of puzzles across many categories: by theme, by popularity, by newest additions.
• Puzzle creation: upload your own image and convert it into a jigsaw for others to play.
• Community features: user profiles, comments, puzzle creation stats, and saving your progress when logged in.
• Daily curated puzzles for fresh content each day.

Use Case Magic

• Great for relaxing with a digital jigsaw when you don’t have a physical one.
• Teachers or parents can use user-created puzzles for fun in classroom or home settings.
• Creative users upload meaningful photos (travel, family, art) and challenge friends to solve them.
• Ideal for casual play, short breaks, and building pattern-recognition and focus skills.

Why it stands out

Unlike standalone puzzle apps that recycle fixed jigsaws, Jigidi taps into a live community generation model—new puzzles are continuously added, and you can contribute your own. The blend of creation + solving + sharing makes it more social and dynamic than many generic jigsaw sites.

---

Masterpieces from the Hive 🎨

Here’s what our community has been creating this week using the Hive Wall:

​

​

32 North Gould Street, Sheridan, Wyoming 82801
​Unsubscribe · Preferences​