TechCrunch Disrupt’s 20th anniversary showcased something rare in tech: ideas that do more than chase hype. I left with a clear view. Real, useful AI is arriving, and it’s time to prioritize what actually improves daily life. The show floor offered plenty of spectacle, but a handful of teams showed how AI can cut waste, ease workloads, and fix broken systems. That is the future I want funded.
The Core Argument
We should back AI that solves concrete problems, not just clever demos. From kitchens to farms to freight, several teams presented systems that earn their keep. The speaker I followed through the event was animated by the same filter. If a product saves time, reduces risk, or lowers costs at scale, it matters. If not, pass.
“We have a fully robotic Korean barbecue restaurant that can cook anywhere, anytime, redefining what dining even means,” said Tor from Shinstar.
“Our north star is productivity… making [developers] super productive but then also giving them that joy while they are coding,” said Arun Gupta of JetBrains.
Those quotes signal a shift: less flash, more function.
Evidence From the Floor
Shinstar packed an autonomous, AI-driven kitchen into a truck. It cooks nonstop, trims labor by about 80%, and is headed for airports. No lines, no wait, just hot food on schedule.
Asleep AI offered contactless sleep tracking with a phone mic. No rings, no straps. Their model reads breathing patterns to map sleep stages and already ships through partners like Samsung and LG. That is usable health tech, not a gimmick.
On the developer side, JetBrains showed its Juny coding agent and announced a public benchmark: the Developer Productivity AI Arena. Claims are cheap; measured results are not.
“What is your next 25.2 miles going to look like? Is V code ready for that? Not right now,” Gupta said, cutting through marketing fog.
Dennis Sheriive, JetBrains’ AI group product manager, stressed trust and user control. That stance matters in an era of unchecked automation.
Glide, the competition winner, built a truck that turns into a train in under 90 seconds. It carries up to 40,000 pounds, runs electric and autonomous, and uses its Ezra 16 system to route trips. America uses about 30% of rail capacity; Glide wants to push that dramatically higher. That means fewer semis, lower emissions, and faster freight.
Two farming startups showed the same practical streak. TensorField Agriculture’s robots identify weeds and torch them with superheated vegetable oil, cutting chemicals and crews. Instacrops tunes irrigation for each field, helping farms use roughly 30% less water while raising profits about 12%. Uni Bio even turned shrimp shells into a nano-powder that helps plants absorb nutrients and pesticides, potentially halving chemical use while lifting yields.
Short List: What Deserves Fast-Track Support
- AI that removes friction for workers and customers.
- Tools measured by open benchmarks, not slide decks.
- Systems that reduce waste: time, water, chemicals, and fuel.
- Infrastructure plays with clear incentives for users and partners.
These aren’t shiny toys. They are upgrades to daily systems.
The Outliers And The Skeptic In Me
General Neuro’s brain-tingling headset was bold and fun to watch. I’m not ready to crown it as learning’s missing link. Mild electrical stimulation has research behind it, but translation to consumer gains needs careful proof and safety data. If you pitch brain boosts, bring peer-reviewed results and guardrails.
Gargoyle Systems’ drone detection network also raised questions. Paying people to host nodes is clever, and the use case is real for airports and stadiums. But local rules, data accuracy, and maintenance will decide whether the grid scales cleanly.
What Should Happen Next
Investors should reward measurable outcomes over performative demos. Back teams that ship, test, and publish results. Cities and agencies should pilot freight mode-shifting and drone visibility in targeted corridors. Developers should rally around public benchmarks like JetBrains’ to sort real gains from noise.
I left convinced: the useful AI wave is here. It cooks, it irrigates, it ships containers, and it respects the worker at the center. Let’s fund the builders who make life simpler, safer, and cleaner—and hold them to proof, not promises.
Frequently Asked Questions
Q: Which ideas from the event felt ready for real use?
Shinstar’s robotic kitchens, Asleep AI’s contactless sleep tracking, Instacrops’ irrigation guidance, and Glide’s road-to-rail freight system looked closest to delivering near-term value.
Q: Why highlight JetBrains’ benchmarking effort?
Public, repeatable tests separate marketing claims from measurable productivity. A shared scoreboard helps teams choose tools that actually save time.
Q: Is the brain-stimulation headset safe and effective?
The concept has research roots, but consumer results and safety standards need rigorous, transparent validation before everyday adoption makes sense.
Q: How could Glide change freight movement?
By shifting cargo from roads to underused rail, it could cut traffic and emissions while improving reliability. The key is scaling operations and partnerships with rail operators.
Q: What matters most when judging new AI products?
Look for clear outcomes, user control, open metrics, and practical savings. If a tool reduces waste or risk at scale, it belongs on the shortlist.
