它從嘈雜的桌子開始

這是在2015年。研究人員是本·菲爾丁（Ben Fielding），他建造了一台裝有早期GPU的大型機器來開發AI。

A noisy desk in a lab at Northumbria University, in northern England, is where a young AI researcher began his PhD track in 2015. He was working on some unorthodox ideas, exploring how “swarms” of AIs could talk to each other and learn from each other to improve the collective whole. But he was handcuffed by the realities of that noisy machine, which he had built to house early GPUs for his research.

在英格蘭北部諾森比亞大學的一家實驗室中的一間嘈雜的辦公桌是一位年輕的AI研究員在2015年開始他的博士學位曲目的地方。他正在研究一些非正統的想法，探索AIS的“群體”如何可以互相交談並互相學習並相互學習以改善集體整體。但是，他被那台嘈雜的機器的現實戴上了手銬，這是他為研究的早期GPU所建造的。

The machine was so loud it annoyed his lab-mates, and he had crammed it beneath his desk, leaving no room for his legs, which he had to awkwardly stick to the side of the desk.

這台機器很大，使他的實驗室煩惱，他把它塞在桌子下面，沒有腿部的空間，他不得不尷尬地貼在桌子的側面。

At the time, Google was doing similar research with thousands of GPUs in a data center.

當時，Google正在數據中心對數千個GPU進行類似的研究。

“The things they were doing weren't crazy. I knew the methods … I had lots of proposals, but I couldn't run them,” says Ben Fielding, who is now CEO of Gensyn and a speaker at Consensus 2025.

“他們所做的事情並不瘋狂。我知道這些方法……我有很多建議，但我無法運行它們。”現任Gensyn首席執行官，共識2025年的發言人Ben Fielding說。

Fielding and his co-founder, Harry Grieve, started Gensyn in 2020, with the goal of creating a decentralized network for machine intelligence.

Fielding和他的聯合創始人Harry Grieve於2020年創立了Gensyn，目的是為機器智能創建分散的網絡。

Gensyn is now rolling out the first stage of its Testnet, and the company's early tools are beginning to trickle out into the wild. Gensyn recently released its “RL Swarms” protocol, a descendant of Fielding's PhD work, and just launched Testnet, which brings blockchain into the fold.

Gensyn現在正在其測試網的第一階段推出，該公司的早期工具開始滴入野外。 Gensyn最近發布了其“ RL Swarms”協議，該協議是Fielding博士學位作品的後代，並剛剛啟動了TestNet，該協議將區塊鏈帶入了折疊。

Gensyn's Testnet is a beta version of the company's core infrastructure, offering a glimpse into the broader vision.

Gensyn的TestNet是該公司核心基礎架構的Beta版本，可瞥見更廣闊的視野。

In this interview leading up to the AI Summit, at Consensus in Toronto, Fielding gives a primer on AI Swarms, explains how blockchain snaps into the puzzle, and shares why all innovators — not just tech giants — “should have the right to build machine learning technologies.”

在這次接觸到AI峰會之前，在多倫多共識上，Fielding對AI群進行了入門，解釋了區塊鏈如何進入難題，並分享了為什麼所有創新者（不僅僅是技術巨頭）都應該有權建立機器學習技術。 ”

This interview has been condensed and lightly edited for clarity.

這次採訪已被凝結和輕微編輯，以澄清。

Jeff Wilser: Gensyn just launched its Testnet. What’s the gist of what it is?

傑夫·威爾瑟（Jeff Wilser）：Gensyn剛剛推出了Testnet。它是什麼要點？

Ben Fielding: It's the addition of the first MVP features of blockchain integration with what we've launched so far. So a few weeks ago, we launched RL Swarm, which is reinforcement learning, post-training as a peer-to-peer network.

本·菲爾丁（Ben Fielding）：這是區塊鏈集成的第一個MVP功能與我們迄今為止推出的功能。因此，幾週前，我們啟動了RL Swarm，這是加強學習，作為對等網絡進行培訓。

The easiest way to think about it is when a pre-trained model goes through reasoning training — like Deep-R1 — it learns to critique its own thinking and recursively improve against the task. It can then improve its own answer.

思考它的最簡單方法是，當預先訓練的模型通過推理培訓（例如Deep-R1）學習時，它學會了批評自己的思維，並遞歸地反對任務。然後可以改善自己的答案。

We take that process one step further and say, "It's great for models to critique their own thinking and recursively improve. What if they can talk to other models and critique each other's thinking?" If you get many models together in a group that can all talk to each other, they can start learning how to send information to the other models … with the overall goal of improving the entire swarm itself.

我們將這一過程進一步邁出了一步，並說：“模型可以批評自己的思維並遞歸改進。如果他們可以與其他模型交談並批評彼此的思想，該怎麼辦？”如果您在一個可以互相交談的小組中將許多模型聚在一起，他們可以開始學習如何將信息發送給其他模型……總體目標是改善整個群體本身。

So that's the swarm training method, which allows many models to combine in parallel to improve the outcome of a final meta-model that you could create from those models. But at the same time, you have every single individual model just improving on its own. So if you were to come along with a model on a MacBook, join a swarm for an hour and then drop back out again, you would have an improved local model based on the knowledge in the swarm, and you would have also improved the other models in the swarm. It's this collaborative training process that any model can join and any model can do. So that's what RL Swarm is.

因此，這就是群訓練方法，它允許許多模型並行組合，以改善您可以從這些模型中創建的最終元模型的結果。但是與此同時，您只有單獨的單獨模型只是自行改進。因此，如果您要在MacBook上配備一個模型，請加入一群小時，然後再次退出，您將根據群體的知識進行改進的本地模型，並且您還可以改善群中的其他模型。這是任何模型都可以加入的協作培訓過程，任何模型都可以做到。這就是RL群。

Wilser: Okay, and then where does blockchain come in?

威爾瑟：好的，然后區塊鏈進來？

Ben Fielding: So the blockchain is us moving forward some of the lower-level primitives into the system.

本·菲爾丁（Ben Fielding）：因此，區塊鍊是我們向系統推進系統的一些低級原始圖。

Let's just pretend that someone doesn't understand the phrase "lower-level primitives." What do you mean by that?

讓我們假裝某人不了解“低級原語”一詞。那是什麼意思？

If you think about the software stack, you've got a GPU stack in a data center. You've got drivers on top of the GPU. You've got operating systems, virtual machines. You've got all this stuff going up.

如果您考慮軟件堆棧，則數據中心有GPU堆棧。您已經有驅動程序在GPU之上。您有操作系統，虛擬機。您已經上了所有這些東西。

A lower-level primitive is the closest to the bottom foundation in the tech stack.

一個低級原始的原始性是技術堆棧中最接近的基礎。

Wilser: Yes, exactly. And the RL Swarm is a demonstration of what's possible, basically. It's just a somewhat hacky demo of doing really interesting large-scale, scalable machine learning. But what Gensyn's been doing for the past four-plus years, realistically, is building infrastructure. And so we're in this period now where the infrastructure is all at that v0.1 sort of beta level. It's all done. It's ready to go. We have to figure out how to show the world what's possible when it's quite a big shift to the way people think of machine learning.

威爾瑟：是的，正是。基本上，RL群是可能的。這只是做一個非常有趣的大規模，可擴展的機器學習的刺激性演示。但是，實際上，Gensyn在過去的四年中一直在建立基礎設施。因此，我們現在正處於基礎架構處於v0.1 beta級別的時期。一切都完成了。準備就緒了。我們必須弄清楚當人們對機器學習的看法發生很大的轉變時，如何向世界展示可能會有什麼可能。

It sounds like you guys are

聽起來你們是