STARTUP X
The F42 brief #034
Here’s your Monday dose of The F42 Brief.
📈 Trending Now
The one thing that’s making all the noise …. with a bit of fact checking.
💡Innovator Spotlight
Founders that are doing stuff differently.
🛠️ Tool of the Week
Something useful to up your startups velocity.
📌 Note to Self
Stuff I constantly remind myself about, don’t want make the same mistakes again.
📈 Trending Now
Apple Cuts Through the AI Hype and the Crap
😩 I’ve grown a bit tired of all the AI hoopla—the grand claims of “AGI” from the usual suspects. Here’s the blunt truth Apple lays out in its “Illusion of Thinking” paper:
🚫 GenAI systems can’t reason.
🤖 They aren’t intelligent.
🎲 They’re just guessing.
🧠❌ They don’t think—and never will.
🏗️ The whole premise rests on false assumptions.
What Apple did was beautifully straightforward 🎩: they built four classic puzzles—Tower of Hanoi 🏰, checker-jumping ⏩, river-crossing 🌊 and blocks world 🧱—with strictly controlled difficulty and zero data leakage. Then, under the same compute budget, they pitched “chain-of-thought” models against plain LLMs. The outcome? Brutal clarity:
📉 Low complexity: vanilla models beat the “thinking” ones.
⚖️ Medium complexity: chain-of-thought grabs a modest edge.
💥 High complexity: both dive to zero—push them past basics and they flatline.
Even more telling: as tasks grow thorny, the so-called reasoning models simply slack off, despite having tokens to spare—an almost comical scaling failure 🙈.
In plain English, these systems don’t solve new problems; they regurgitate patterns they’ve seen 🔍. Strip away clever prompts and memorised answers, and their “reasoning” evaporates. Far from sentient sidekicks, they’re glorified autocomplete engines.
The hard truth for founders, framers and funders? Don’t fall for the myth of thinking machines 🛑. Treat GenAI as a razor-sharp prediction tool, not an oracle 🔮. Build with your eyes wide open and strategies grounded in reality—because when it comes to genuine multi-step reasoning, today’s tech just can’t cut it.
🙏 And perhaps that’s fitting: if we humans often lean on pattern-matching too, maybe these systems are just holding up a mirror to our own shortcuts. Until genuine reasoning arrives, a dash of blind faith remains our most human trait.
The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models via the Lens of Problem Complexity
💸 New Funds looking to give you cash
Look out for the The Raise Report this Wednesday, bring you all the new funds that have raised capital and looking to deploy.
Landing in Dubai –
What Global Startups Need to Know
With our Partner - Dubai Technology Entrepreneur Campus (Dtec).
Thinking about expanding into Dubai or the broader GCC region? Join this high-impact session designed to give global startups the insight, access, and tactical guidance needed to land, launch, and grow in the UAE.Hear directly from public and private leaders on incentives, setup options, the investor and tech ecosystem, and the reality of scaling in one of the world’s fastest-growing innovation hubs.
💡Innovator Spotlight
Last week I listened to “AI × Crypto” episode of Web3 with A16Z Crypto, and let me tell you—it’s blockchain’s moment in the sun.
🔒 Trust & Identity, Supercharged
I’ve seen blockchains keep data honest, but now AI is sniffing out on-chain shenanigans—front-running bots, wash trades, you name it. And with deepfakes everywhere, Proof of Personhood protocols (AI vision plus zero-knowledge proofs) finally nail “you are you” on-chain. That’s foundational for any permissionless network.🤖 Autonomous Agents Go Public
Those demo “robot traders” you saw in labs? They’re live now. Imagine an AI that rebalances your DeFi positions, snags yield and locks in gains—no human needed. In gaming, in-world bots will earn and spend tokenised assets you truly own. Suddenly, blockchains aren’t just ledgers; they’re the runtime for AI economies.🌐 DePIN Hits the Real World
Decentralised Physical Infrastructure Networks are getting brains. From peer-to-peer internet to EV chargers, edge AI feeds validated data into smart contracts. Your car could book a charger, pay in crypto and update the network—autonomously. That’s machines transacting with machines, and it’s happening today.🔄 Tokenised Intelligence
Think Uniswap for AI models and data sets: anyone can mint, license or vote on algorithms via tokens. It flips AI from a corporate black box into a community-driven asset—and that’s huge for decentralisation.⚖️ The Balancing Act
Expect hybrids—heavy-duty AI off-chain with on-chain proofs for transparency. Regulators will want audit trails; blockchains deliver. The protocols that nail this choreography will own the next internet wave.
Why It Matters
By 2025, we won’t just store data on blockchains—we’ll run intelligent, self-optimising networks on them. If you’re not building with both AI smarts and crypto trust, you’re missing the tidal wave. This is blockchain’s golden age, and it’s happening right now.
Catch the episode here: AI x Crypto
If you have not joined the Fusion42 Community on Telegram — A space for Founders, Framers, and Funders who show up to get **it done,
it is probably time to do so.
For the ❤️ of Startups
🛠️ Tools of the Week
The Hottest AI x Blockchain Platforms to Watch in 2025
AI and blockchain are a must-watch in 2025 because together they deliver unprecedented data security, transparency, and user control—enabling trustworthy, privacy-preserving analytics and automation at scale
Fetch.ai
Autonomous AI agents for automated, trustless transactions and data sharing across networks.
SingularityNET
Decentralized marketplace for AI algorithms—build, share, and monetize AI services on-chain.
Ocean Protocol
Tokenized data marketplace—securely share, monetize, and analyze data using AI models on blockchain.
ChainGPT
AI for smart contract generation, auditing, and technical analysis—plus AI-generated NFTs and agent infrastructure.
Cortex
Enables AI model execution directly on-chain, bringing machine learning to smart contracts.
Render Network
Decentralized GPU and AI compute marketplace—power AI/ML tasks and rendering with idle resources.
The Graph
Decentralized protocol for indexing and querying blockchain data using AI-powered subgraphs.
Artificial Superintelligence Alliance (ASI)
Universal AI token (Fetch.ai, SingularityNET, Ocean) for secure, decentralized AI services and interoperability.
Griffain
AI agent coordination on Solana for complex on-chain actions and automation.
Augur
Decentralized prediction markets using AI for outcome verification and market creation.
Agent-based automation:
Platforms like Fetch.ai and Griffain use AI agents to automate on-chain tasks, creating new possibilities for dApps and DeFi.
Decentralized AI marketplaces: SingularityNET and Ocean Protocol let you build, deploy, and monetize AI models and data securely.
On-chain AI execution: Cortex and ChainGPT bring direct AI computation and smart contract generation to the blockchain layer.
Compute and data marketplaces: Render Network and The Graph unlock decentralized compute and data analytics, vital for AI/ML and transparency.
Universal AI tokens: The ASI alliance is pioneering cross-platform AI services and interoperability.
AI-powered prediction and compliance: Augur and Ocean Protocol are pushing boundaries in prediction markets and privacy-preserving compliance.
📌 Note to Self
Thank you for reading. If you liked it, share it with your friends, colleagues and everyone interested in the startup Investor ecosystem.
If you've got suggestions, an article, research, your tech stack, or a job listing you want featured, just let me know! I'm keen to include it in the upcoming edition.
Please let me know what you think of it, love a feedback loop 🙏🏼
🛑 Get a different job.
Subscribe below and follow me on LinkedIn or Twitter to never miss an update.
For the ❤️ of startups
✌🏼 & 💙
Derek