I asked an AI swarm to fill out a March Madness bracket — here’s what happened

Imagine that a large team of 200 people, brainstorm ideas, knowledge, the controversial alternatives, and could make a thoughtful real-time negotiation in the AI-optimized solutions. This is possible – and if it is, they are strengthening collective intelligence?

A new generative AI technology, scratching intelligence (or simply hyperkat) It allows you to deal with real-time talks, real-time negotiations in any size and unite on AI-optimized solutions quickly. In order to put this, in the sense, I asked the EU research group to bring together 50 casual sports fans and quickly creating a March dangness bracket with a majority bracket.

Prior to how the experience goes, I should explain why we can’t bring only 50 people to an enlargement meeting and explain them to create a bracket quickly. Studies show the ideal measurement for productive real-time conversation Only 4-7 people. In small groups, each person takes a good ether time to express his views and has a low waiting time to answer others. However, as the group size grows in size, the waiting period increases and becomes a number of monologies by several people. More than 20 people, it is chaos.

So how can 50 people spend a conversation, or 250, even 2500 people?

Hyperchat works by breaking any large group to a Parallel subgroup set. Later, people withdrawing human concepts in the local group, these concepts add these concepts to each sub-group called “Talk Surrogat”, which quickly shares as a natural dialogue box. These surrogate agents allow all sub-groups to overlap, local conversations to weave a large conversation. And it works for brainstorming, to activate any size groups, prioritize, dispute and unite in real time mode.

Hyperchat technology has been invented significantly not to communicate and cooperate on a large scale Strengthen group exploration. Progress was fast on this front and already uses the commercial platform for enterprise groups called Thought^® This allows hundreds of people to be optimized in real time.

But hyperchat technology really makes teams smarter (and can predict the results of March madness)?

In order to test this in full public appearance, I thought the team in the unequivocal EU, I thought 50 casual sports fans and March to create madness brackets. Entered after brackets that arise ESPN March Madness Competition So we can watch how good 30 million other people are. The parentheseth ESPN competition, created by 50 random people, has a 99th percentage (top 1.4%). Here’s the stats:

Of course, the tournament lasts this week, but so far the teams of intelligence created among the group of supporters who are talking about this hyper. This technology is not the first time I am surprised.

One 2024 Education Carnegie Mellon and Safigali EU researchers were asked to accept standard IQ tests from 35 groups by hyperchat. The results showed that an effective IQ (97th percentage) effective IQ (97th interest) is effective in the average (50th percentage) in the average (50th percentage) in the hyperchat platform while self-employed. This is a gift level performance.

In the other 2024 Education75 groups of groups were asked to brainstorm real-time to solve a creative problem. These groups have repeatedly used the standard chat and used the view of the thinking using hyperkat. The groups later compared the practice and communicated with hyperchat, they felt more productive, more collaborated and better solutions (P <0.001). The solutions that arise in the process and "more ownership" were more likely to "buy" (P <0.001).

This technology excited me for a long time, not because it makes human groups smarter. It also has the ability to break hybrid groups of human participants and AI agents, allows you to cooperate on an unlimited scale that allows optimized decisions keep people in the loop. This requires the addition of a second type of AI agent to the hyperchat structure, known as the “Contributor Agent”. These agents provide factual content in real time to support the discussion of the ongoing person. The goal is to ensure a collective collective management of hybrid.

This hybrid technique was first tested 2024 Education This brought together groups of human and AI agents and AI agents using hyperchat structure in real time. The results showed that major cooperation teams were a high productive tool for the observation of hyperchat, and 87% of the participants have caused better decisions in this regard.

In general Talking defective intelligence In real mode, the strong use of AI agents that can radically change cooperation Potential conversations with any size. Given that there are more than 30,000 employees of 1000 companies in the average Fortune 1000, it can solve a long-term buttleneck with absolute controversy in small teams. While keeping people in control, AI is an effective way to use the power in critical decisions.

Men’s march the madness tournament lasts this week. Something can happen, but I suspect that collective intelligence used by these 50 casual sports fans will work very well. We’ll see …

Louis found Rosenberg Immersion Corp and The wrong ai.