Jeff (Startup Whisperer)

The most successful field in computer science right now is also the most anxious. You can feel it in Reddit threads, conference hallways, and DMs: something about how we do ML research is off. The pace is intoxicating, the progress is real—and yet the people building it are quietly asking, “Is this sustainable? Is this still science?”

That tension is the story: a field that went from scrappy outsider to global infrastructure so fast it never upgraded its operating system. Now the bugs are showing.

When “More Papers” Stops Feeling Like Progress

In theory, more research means more discovery. In practice, we’ve hit the point where conference submission graphs look like someone mis-set the y-axis. Flagship venues are drowning in tens of thousands of papers a year, forcing brutal early rejections and weird hacks to keep the system from collapsing.​

From the outside, it looks like abundance. From the inside, it feels like spam. Authors optimize for “accepted somewhere, anywhere” instead of “is this result robust and useful?” Reviewers are buried. Organizers are pushed into warehouse logistics instead of deep curation. The whole thing starts to feel like a metrics game, not a knowledge engine.​

When accepted papers with solid scores get dropped because there isn’t enough physical space at the venue, that’s not a nice problem to have. That’s a signal the model is mis-specified.​

Quality Debt and the Reproducibility Hangover

Meanwhile, a quieter crisis has been compounding: reproducibility. Code not released. Data not shared. Baselines mis-implemented. Benchmarks overfit. Half the field has a story about trying to re-run a “state of the art” paper and giving up after a week.​

This isn’t just a paperwork problem. If others can’t reproduce your result:

• No one knows if your idea generalizes.

• Downstream work might be building on a mirage.

• Real-world teams burn time and budget chasing ghosts.

As models move into medicine, finance, and public policy, “it sort of worked on this dataset in our lab” is not a pass. Trust in the science behind ML becomes a hard constraint, not a nice-to-have.​

Incentives: Optimizing the Wrong Objective

Zoom out, and a pattern appears: the system is rewarding the wrong things.

• Novelty over reliability.

• Benchmarks over messy, real problems.

• Velocity over understanding.

The fastest way to survive in this game is to slice your work into as many publishable units as possible, push to every major conference, and pray the review lottery hits at least once. Deep, slow, high-risk ideas don’t fit neatly into that cadence.

And then there’s the talent flow. The best people are heavily pulled into industry labs with bigger checks and bigger GPUs. Academia becomes more about paper throughput on limited resources. The result: the people with the most time to think have the least compute, and the people with the most compute are often on product timelines. Misalignment everywhere.​

The Field’s Growing Self‑Doubt (That’s Actually Healthy)

Here’s the twist: this wave of self-critique is not a sign ML is dying. It’s a sign the immune system is finally kicking in.

Researchers are openly asking:

• Are we publishing too much, learning too little?

• Are our benchmarks telling us anything real?

• Are we building tools that transfer beyond leaderboards into the world?

When people who benefit from the current system start calling it broken, pay attention. That’s not nihilism; that’s care. It’s a field realizing it grew up faster than its institutions did—and deciding to fix that before an AI winter or an external backlash does it for them.​

What a Healthier ML Research Culture Could Look Like

If you strip away the institutional inertia, the fixes aren’t mysterious. They’re the research equivalent of “stop pretending the plan is working; start iterating on the process.”

Some levers worth pulling:

• Less worship of novelty, more respect for rigor. Make “solid, careful, negative-result-rich” a first-class contribution, not a consolation prize.

• Mandatory openness. If it can be open-sourced, it should be. Code, data, evaluation scripts. No artifacts, no big claims.

• Different tracks, different values. Separate venues or tracks for (a) theory, (b) benchmarks, (c) applications. Judge each by the right metric instead of forcing everything through the same novelty filter.

• Incentives that outlast a deadline. Promotion, funding, and prestige that factor in impact over time, not just conference logos on a CV.

None of this is romantic. It’s plumbing. But if you get the plumbing right, the next decade of ML feels very different: fewer hype cycles, fewer brittle “breakthroughs,” more compounding, reliable progress.

If You’re an ML Researcher, Here’s the Move

You can’t fix the whole ecosystem alone—but you can run a different local policy.

• Treat your own beliefs like models: version them, stress-test them, deprecate them.

• Aim for “someone else can reproduce this without emailing me” as a hard requirement, not an aspiration.

• Choose questions that would matter even if they never hit a top-tier conference.

• Remember that “I don’t know yet” and “we couldn’t replicate it” are signs of seriousness, not weakness.

Machine learning isn’t in crisis because it’s failing. It’s in crisis because it’s succeeding faster than its institutions can adapt. The people who will matter most in the next decade aren’t the ones who ride this wave blindly—they’re the ones who help the field course-correct in public, with less ego and more evidence.

World models are quietly transforming AI from text predictors into systems that understand and simulate the real world. Unlike large language models (LLMs) that predict the next word, world models build internal representations of how environments evolve over time and how actions change states. This leap from language to spatial intelligence promises to unlock AI capable of perceiving, reasoning, and interacting with complex 3D spaces.

Fei-Fei Li calls world models "the next frontier of AI," emphasizing spatial intelligence as essential for machines to see and act in the world. Yann LeCun echoes this urgency, arguing that learning accurate world models is key to human-level AI. His approach highlights the need for self-supervised learning architectures that predict world states in compressed representations rather than raw pixels, optimizing efficiency and generalization.

Leading efforts diverge into three camps. OpenAI’s Sora uses video generation transformers to simulate physical environments, showing emergent long-range coherence and object permanence, crucial for world simulation. Meta’s Joint Embedding Predictive Architecture (V-JEPA) models latent representations of videos and robotic interactions to reduce computational waste and improve reasoning. Fei-Fei Li’s World Labs blends multimodal inputs into spatially consistent, editable 3D worlds via Marble, targeting interactive virtual environment generation.

The commercial potential is looking to be enormous. Over $2 billion was invested across 15+ world model startups in 2024, with estimates valuing the full market north of $100 trillion if AI masters physical intelligence. Robotics leads near-term value: enabling robots to safely navigate unstructured environments requires world models to predict object interactions and plan multi-step tasks. NVIDIA’s Cosmos infrastructure accelerates physical AI training with synthetic photorealistic data, while companies like Skild AI have raised billions by building massive robotic interaction datasets.​

Autonomous vehicles also tap world models to simulate traffic and rare scenarios at scale, cutting down expensive on-road tests and improving safety. Companies like Wayve and Waabi leverage virtual worlds for pre-labeling and scenario generation, critical in achieving full autonomy. Meanwhile, the gaming and entertainment sector is the most mature commercial playground, with startups using world models to generate dynamic game worlds and personalized content that attract millions of users almost overnight

Specialized industrial applications—engineering simulations, healthcare, city planning—show clear revenue pathways with fewer competitors. PhysicsX’s quantum leap in simulation speed exemplifies how tailored world models can revolutionize verticals where traditional methods falter. Healthcare and urban planning stand to gain precision interventions and predictive modeling unparalleled by current AI.

The funding landscape reveals the importance of founder pedigree and scale. Fei-Fei Li’s World Labs hit unicorn status swiftly with $230 million raised, Luma AI secured $900 million Series C for supercluster-scale training, and Skild AI amassed over $1.5 billion focused on robotics. NVIDIA, while a supplier, remains a kingmaker, providing hardware, software, and foundational models as a platform layer—both opportunity and competition for startups.

Crucially, despite staggering investment, gaps abound—technical, commercial, and strategic. Training world models requires vast, complex multimodal datasets rarely available openly, creating defensive moats for data-rich startups. Models still struggle with physics accuracy, generalization to novel scenarios, and real-time performance needed for robotics or autonomous vehicles. Startups innovating around efficiency, transfer learning, sim-to-real gaps, and safety validation have outsized opportunities.

On the market front, vertical-specific solutions in healthcare, logistics, and defense are underserved, offering fertile ground for founders with domain expertise. Productizing world models requires bridging the gap from lab prototypes to robust, scalable deployments, including integration tooling and certification for safety-critical applications. Startups enabling high-fidelity synthetic data generation are becoming ecosystem enablers.​

Strategically, founders must navigate open research—like Meta’s V-JEPA—and proprietary plays exemplified by World Labs. Standardization and interoperability remain open questions critical for ecosystem growth. Handling rare edge cases and ensuring reliable sim-to-real transfer are gating factors for robotic and autonomous systems.

For investors, the thesis is clear but nuanced. Robotics world models, vertical AI for high-value industries, infrastructure and tooling layers, and gaming are high-conviction bets offering a blend of risk and clear pathways to market. Foundational model companies with massive compute and data moats present risky but lucrative opportunities, demanding large capital and specialized talent. Efficiency, differentiated data, and agile product-market fit matter more than raw scale alone.

The next 24 months will crystallize market winners as world models shift from research curiosity to mission-critical AI infrastructure. Founders displaying relentless adaptability, technical depth, and deep domain insight will lead the charge. Investors who balance bets across foundation layers and vertical applications, while embracing geographic and stage diversity, stand to capture disproportionate value.

While the industry watches language models, the less flashy but more profound revolution is unfolding quietly in world models—systems that don’t just process language but build a mental map of reality itself. These systems will define the next era of AI, shaping how machines perceive, interact, and augment the physical world for decades.

That’s the state of play. The winners will be those who combine technical innovation with pragmatic business sense, and above all, a ruthlessly adaptive mindset to pivot rapidly as the frontier evolves.

Build global, or get boxed in. Singapore is an exceptional launchpad for deep tech—world-class research, predictable regulation, dense talent, and brand equity that travels—but the world won’t bend to our advantages unless the execution is ruthless, market-led, and globally capitalized. The playbook is simple to say, hard to do: prove your science is best-in-class, lock real customer pain with a sharp ICP, market like a category winner to reach specialist deep tech capital, and hire a killer commercial bench through a global search. Do these in parallel, not sequence.​

Start with the truth: your research must actually be world-leading. Not locally/regionally excellent—globally defensible. Strong patent estates correlate with outlier outcomes because patents aren’t just legal armor; they are signals of technical scarcity, negotiation leverage, and acquisition currency. In Europe, deep tech unicorns carry dramatically larger patent portfolios than general tech peers, and the same pattern holds across AI hardware, robotics, and biotech. If your tech wins only in the lab, you don’t have a moat—you have a demo. File early and internationally via PCT, cover where competitors operate, and budget real money for freedom-to-operate and continuations; it’s the price of building in hard tech. Then pressure test the science in public: publish, present, and partner with tier‑one labs. NUS’s new co‑investment flywheel and Stanford collaboration are the right instincts—cross‑border validation tightens the BS filter and compounds credibility with buyers and investors.​

Next, stop letting tech chase the market. The fastest way to die in deep tech is mistaking novelty for need. Traditional PMF heuristics mislead here; what you need is technology‑market fit: a specific workflow, buyer, and willingness‑to‑pay that your product makes meaningfully better under real constraints (regulatory, reliability, integration). Work the TRL stack with intent: at TRL 1–4, mine “earned secrets” from the field before you write code; at TRL 4–6, validate multi‑stakeholder adoption (clinical, compliance, procurement); at TRL 7–9, convert pilots into lighthouse accounts with signed commercial terms, not vibes. Precision beats ambition: define a sharp ICP (role, budget, system dependencies, success metric) and a wedge (one or two killer workflows) that lands ROI in <90 days. Remember the graveyard: Lilium and Arrival raised billions and still cratered—multi‑front innovation without a narrowed use-case and industrial discipline is how you burn years and trust.​

Now the uncomfortable part: you must be loud—and surgical—about your story to attract the right capital. Southeast Asia doesn’t have enough patient deep tech funding to carry you through multiple cycles; winning requires a global investor map and a narrative that decodes risk for them. The good news: specialized capital is abundant and hunting—Europe alone pushed ~€15B into deep tech in 2024; AI is swallowing the lion’s share of global deal value; Switzerland allocates a majority of VC to deep tech. But visibility is earned. Use credibility magnets: international conferences and trades shows for global stage time, third‑party validation, and platform grants; institutional tie‑ups to signal momentum; university venture programs to anchor de‑risked spinouts. Ditch feature‑speak. Lead with outcomes: “cut false positives 30%, lifted yield 12%, reduced cost per cycle 40%,” tied to buyer P&L. Then make the moat explicit—IP, data exclusivity, regulatory posture, and integrations that raise rip‑and‑replace costs.​

Talent is the force multiplier. Technical founders don’t have to become CROs—but someone elite must own revenue, sequencing, and global expansion. Industrial-grade deep tech fails not because of bad science, but because management, manufacturing, and GTM never catch up to the physics. Time the hires. Pre‑PMF: keep the founder selling; add a solutions lead who speaks both code and plant floor. Approaching scale: bring in a CRO/CCO with credible enterprise cycles in your domain; under ~$2M ARR, hiring a full CRO is usually premature—prove repeatability first. Don’t local‑shop the search. Run a retained, global process: firms with deep tech benches can screen for dual fluency (technical rigor + enterprise sales), access passive candidates, and de‑risk culture/comp plans across geos. Yes, it’s expensive. A bad executive hire costs more.​

Finally, design for global from day one. Keep R&D in Singapore for cost, quality, and IP control; put GTM leadership where the buyers are. Hub‑and‑spoke works: a “customer obsession” pod in the U.S. or EU (seller, SE, product) translating field signal into roadmap; core science and data ops stay home for velocity and security. Start narrow, win deeply—one metro, one buyer, one killer workflow—then expand into adjacencies with proof and references. Use platform distribution early (hyperscaler co‑sell, OEMs, integrators) to compress sales cycles and credibility debt. Make momentum visible: case studies, third‑party benchmarks, security certifications in flight.​

The meta‑skill that ties it all together: adaptive speed. Ego last, evidence first. Install kill‑switch metrics. Run red‑team reviews monthly. Update the narrative when reality changes. Global winners aren’t the ones who never miss; they’re the ones who correct in public, recruit ahead of the curve, and keep the bar on science and outcomes where the world can see it. 

Singapore gives you the runway. The world sets the bar. Make the science undeniable, the market signal unmistakable, the capital global, and the team formidable.