Alexander Long (@AlexanderLong) / X

Alexander Long

523 posts

Alexander Long

@AlexanderLong

Founder @Pluralis | ML PhD

Joined July 2023

Pinned
Alexander Long
@AlexanderLong
Dec 5, 2025
Since I started getting interested in ML I got it in my head that all I wanted to do was one smart thing that I could look back on and be satisfied that I did. Most papers are kinda bad even if they get accepted - the idea is very incremental, or it's just not that good an idea,
19K
Alexander Long
@AlexanderLong
Aug 26, 2024
So crazy to me that 6 months ago literally everyone serious AI researcher I talked to told me stuff like this wasn't feasible. Decentralized Training is gonna work - we're still at DDP stage and there's a long way to go for giant model runs, but no doubt where the vector of
Nous Research
@NousResearch
Aug 26, 2024
What if you could use all the computing power in the world to train a shared, open source AI model? Preliminary report: github.com/NousResearch/D… Nous Research is proud to release a preliminary report on DisTrO (Distributed Training Over-the-Internet) a family of
44K
Alexander Long
@AlexanderLong
May 29, 2025
When I started Pluralis I thought it was going to take 2-3 years and several papers to get to this point. I was considered borderline delusional to hold that view. Well: computational graph itself is split over nodes. Nodes are physically in different continents. 8B model. No
Pluralis Research
@Pluralis
May 29, 2025
Almost exactly 1 year ago today Pluralis set out to solve a very difficult problem. We are pleased to announce an update on that problem. We can now combine small consumer devices, connected via the internet, to train giant models. Full paper release in 72 hours.
31K
Alexander Long
@AlexanderLong
Jul 8, 2024
Decentralized training is much closer to reality than broadly thought. The popular narrative here is wrong. Full detailed article below
69K
Alexander Long
@AlexanderLong
May 5, 2025
Probably biggest week in Decentralized Training to date off back of ICLR and more about to come out. Summary of situation as it stands today: 1. Decentralized RL post-training is clearly working. @gensynai the latest with great results here. This process takes a strong base
17K
Alexander Long
@AlexanderLong
Mar 19, 2025
I never wrote the origin story or motivation for Pluralis down before - but now was a good time to do that. Can read in detail here
Pluralis Research
@Pluralis
Mar 19, 2025
1/ We are excited to announce our $7.6M seed round led by @usv and @coinfund to make true open source AI a reality. Pluralis has a new approach - we're calling this Protocol Learning.
pluralis.ai
Blog | Pluralis Research
Research articles on Protocol Learning, decentralized training, and communication-efficient model parallelism from the Pluralis team.
13K
Alexander Long
@AlexanderLong
Apr 7, 2025
Join to understand how RL post-training actually works
Pluralis Reading Group - DAPO: An Open-Source LLM Reinforcement Learning System at Scale · Luma
From luma.com
6.2K
Alexander Long
@AlexanderLong
Jul 14, 2025
For people not familiar with AI publishing; there are 3 main conferences every year. ICML, ICLR and NeurIPS. These are technical conferences and the equivalent of journals in other disciplines - they are the main publishing venue for AI. The competition to have papers at these
Pluralis Research
@Pluralis
Jul 14, 2025
Pluralis has a main tack paper at ICML this week and the team is in Vancouver running several events. The best will be the Open Source Mixer we're running with @gensynai and @akashnet Thursday night. For anyone interested in decentralised training it should be a great evening.
21K
Alexander Long
@AlexanderLong
Oct 13, 2024
The decentralized training race is now underway which I expect to be the major theme of AI research in 2025, create a new field and kick off a megacycle. What happens once you accept decentralized training is feasible? Pluralis’s answer is: Protocol Models. 1/9
Pluralis Research
@Pluralis
Oct 13, 2024
Article 2: Protocol Learning, Protocol Models and the Great Convergence pluralisresearch.com/p/article-2-pr…
19K
Alexander Long
@AlexanderLong
Apr 18, 2025
The holy grail of decentralized AI is Model-Parallel training over low-bandwidth interconnects. This is what lets you pool small devices over the internet to train giant models. This is all Pluralis cares about.
7.2K
Alexander Long
@AlexanderLong
Mar 19, 2025
Could not be more pumped to have USV involved with Pluralis. When I was making the investor list for the seed round, they were #1.
USV
@usv
Mar 19, 2025
1/ We're excited to announce our investment in @pluralisHQ, advancing "Actually Open AI" through decentralized, collaborative model training. USV and @coinfund co-led the $7.6M seed round, with @variantfund, @topology_vc, and others participating.
6.9K
Alexander Long
@AlexanderLong
Mar 21, 2025
Pluralis is hiring across the board
CoinFund
@coinfund
Mar 21, 2025
The rapid rise of DeepSeek earlier this year provided many developers the tremendous validation that open source AI could overtake centralized machine learning. Still, there’s one nonconsensus approach to training models that is hotly debated in the deAI sphere. CoinFund
00:00
jobs.ashbyhq.com
Pluralis Research Jobs
Pluralis Research Jobs
9.2K
Alexander Long
@AlexanderLong
Apr 30, 2025
We’re publishing three papers (this is the first) addressing core problems in Protocol Learning: multi-participant, low bandwidth model parallel training of foundation models. These papers are workshop papers - they don’t solve the problem (the paper below is in the DP setting) -
Pluralis Research
@Pluralis
Apr 30, 2025
We propose a new Momentum Look-Ahead algorithm at ICLR'25 MCDC which allows heterogenous GPUs to be used with high utilization in decentralized pretraining. Baselines are Async-DiloCo and DyLU which we outperform. blog.pluralis.ai/p/efficient-as…
9.2K
Alexander Long
@AlexanderLong
Sep 23, 2024
I'm completely convinced decentralized training is about to define a mega cycle. There's gonna be massive, ubiquitous real world utility via a convergence of two of the most individually deep technical fields ever... so much of the legit crypto work of the last decade is
8K