Yupp.ai shuts down after raising $33M from a16z crypto’s Chris Dixon
Less than a year after launching, with checks from some of the biggest names in Silicon Valley, crowdsourced AI model feedback startup Yupp.ai is closing its business, the company said Tuesday.
Sometimes an apparently good idea, a big raise from a big-name VC, and a sea of well-connected angel investors is not enough.
Less than a year after launching, Yupp is closing its business, co-founders Pankaj Gupta and Gilad Mishne announced on Tuesday.
Yupp offered a crowdsourced AI model-picking service. It allowed consumers to test and compare results from a supply of 800 AI models for free, including the state-of-the-art ones from OpenAI, Google, and Anthropic. Yupp would return multiple replies from the prompt request, including information or images, and users would offer feedback on which models worked best for them and why.
The idea was to generate anonymized data on what people actually need from AI that the model makers would then pay for. Yupp said it signed up 1.3 million users and collected millions of preferences every month. It even had a leaderboard. The company said it also had a few AI labs as customers.
But alas, it “didn’t reach a strong enough product-market fit” to survive, in part because AI models improved by such leaps and bounds these past few months, the founders said.
While labs are paying big bucks for feedback, the current model — pioneered by companies like Scale AI and Mercor — is to hire specialty experts, like PhDs, and tuck them into the reinforcement learning loop.
On top of that, Silicon Valley is already looking 10 miles down the road, when AI is built for, and being used by, other AIs. Model makers might want some consumer feedback now, but they are largely building for the day when agents, not humans, rule the online world.
Techcrunch event
San Francisco, CA | October 13-15, 2026
“The AI model capability landscape has changed dramatically in the last year alone and will continue to change quickly,” Gupta, Yupp’s CEO, wrote in a post on X about the plans to shutter. “The future is not just models but agentic systems.”
Yupp raised a $33 million seed round in 2024 led by a16z crypto’s Chris Dixon, a giant seed round for its day. In addition, Yupp raised checks from more than 45 angels and small investors, it said. This included luminaries like Google DeepMind chief scientist Jeff Dean; Twitter co-founder Biz Stone; Pinterest co-founder Evan Sharp; and Perplexity CEO Aravind Srinivas.
Gupta said some of Yupp’s employees are joining a “well-known” AI company, and others are looking for their next gig. Yupp did not immediately respond to TechCrunch’s request for comment.
Sign in to highlight and annotate this article

Conversation starters
Daily AI Digest
Get the top 5 AI stories delivered to your inbox every morning.
More about
modellaunchstartupGLM 5.1 vs Minimax 2.7
<!-- SC_OFF --><div class="md"><p>Ok so I've paid for both at their cheapest plans and I have high-level anecdotal feedback on these models.</p> <p><strong>MiniMax 2.7</strong></p> <p>- Extremely Fast</p> <p>- Usage is insane, even at its lowest tier I feel like I could run multiple instances at once without running into session/weekly limits. </p> <p>- Seem to be pivoting themselves into an OpenClaw provider. Their price packges say 'Can power x1 OpenClaw Agent // Can power x2-3 OpenClaw Agents' etc. etc</p> <p>- Not the greatest at understanding codebases and building from scratch. Probably better for smaller tweaks.</p> <p>Overall, I would say this model is worse than Sonnet 4.6 in terms of capability, but price to volume of what you get is absolutely insane, and even its cheapest tier
Caltech Researchers Claim Compression of High-Fidelity AI Models
Article URL: https://www.wsj.com/cio-journal/caltech-researchers-claim-radical-compression-of-high-fidelity-ai-models-e66f31c9 Comments URL: https://news.ycombinator.com/item?id=47593903 Points: 1 # Comments: 0
Knowledge Map
Connected Articles — Knowledge Graph
This article is connected to other articles through shared AI topics and tags.
More in Products

Over 47,000 cabbies adopt Octopus, card readers ahead of e-payment deadline
More than 47,000 taxi drivers have adopted the commercial version of the Octopus app or equipped their cabs with mobile card readers ahead of new electronic payment regulations taking effect in Hong Kong on Wednesday, transport authorities have said. The new measures require cab drivers to provide at least two e-payment options, as part of a broader government push to modernise the industry’s image and service quality. The Transport Department on Tuesday said three dedicated service stations at...

US may reassess Nato ties after Iran war ends, Rubio says
Secretary of State Marco Rubio said the US may need to reassess its relationship with Nato after the Iran war is finished, calling the military alliance’s alleged lack of support during the Middle East conflict “very disappointing”. Rubio assailed Nato members for denying access to military bases, following prior criticism from US President Donald Trump that partners in the security bloc are “cowards” and that the alliance is a “paper tiger”. “The president and our country will have to...

China influencer amasses over 1 million followers in 2 weeks by sharing how to rent a home tips
A Chinese influencer gained more than a million followers in just two weeks by simply teaching people how to rent a home. His efforts earned widespread support and appreciation from Chinese netizens. The influencer, known as Gulumao, was born in the 1990s and is based in Xiamen, Fujian province, southern China. He went viral for offering simple, practical life-skill tutorials, becoming one of the fastest-growing creators on one major platform since the start of this year. On February 26, he...
GLM 5.1 vs Minimax 2.7
<!-- SC_OFF --><div class="md"><p>Ok so I've paid for both at their cheapest plans and I have high-level anecdotal feedback on these models.</p> <p><strong>MiniMax 2.7</strong></p> <p>- Extremely Fast</p> <p>- Usage is insane, even at its lowest tier I feel like I could run multiple instances at once without running into session/weekly limits. </p> <p>- Seem to be pivoting themselves into an OpenClaw provider. Their price packges say 'Can power x1 OpenClaw Agent // Can power x2-3 OpenClaw Agents' etc. etc</p> <p>- Not the greatest at understanding codebases and building from scratch. Probably better for smaller tweaks.</p> <p>Overall, I would say this model is worse than Sonnet 4.6 in terms of capability, but price to volume of what you get is absolutely insane, and even its cheapest tier
Discussion
Sign in to join the discussion
No comments yet — be the first to share your thoughts!