Arjun Karanam

243 posts

Arjun Karanam banner
Arjun Karanam

Arjun Karanam

@QuantumArjun

cofounder at Trajectory (@trajectorylabs) | prev @stanford, @apple

SF Katılım Eylül 2014
821 Takip Edilen1.7K Takipçiler
Sabitlenmiş Tweet
Arjun Karanam
Arjun Karanam@QuantumArjun·
I first met @rronak_ and @MichaelElabd when we were all freshman at Stanford. Today, 8 years later, we’re announcing that we’ve started Trajectory, a research lab and product company building the platform for continual learning. I believe that Continual Learning demands a fundamentally new interface for how we build products. That's a research challenge and a product challenge in equal measure, so we've assembled a team to meet both: researchers from DeepMind, OpenAI, Apple, Meta Superintelligence, Amazon AGI, and Scale AI, and product talent from Stripe and Figma. We’re also partnering with the best AI native companies @Clay, @Harvey, @Decagon, @Mercor, and @RogoAI to power their agentic experiences, and push the boundaries of what agents look like in the real world. Please reach out if you’re excited to build with us!
Ronak Malde@rronak_

Today, @MichaelElabd, @QuantumArjun, and I are excited to announce Trajectory. We are a research lab and product company building the platform for Continual Learning. Our platform unlocks the signal already sitting in product usage, so companies can continuously post-train large-scale agentic models that outperform the frontier. @trajectorylabs We’ve raised $15M from @Conviction, @BessemerVP, @radicalvcfund, @jeffdean, @drfeifei and more. We’re partnering with some of the best AI-native companies: @ClayRunHQ @Harvey, @DecagonAI, @mercor_ai, @RogoAI to power their agentic systems, some of which we are already in production with. We’ve brought together a world class research team from DeepMind, OpenAI, Apple, Meta Superintelligence, Amazon AGI, Scale AI, and an elite product team from Stripe and Figma. AI will never again start on day one. Every correction, every retry, every edit will make products smarter. This is Continual Learning.

English
26
3
85
22.9K
Cody Blakeney
Cody Blakeney@code_star·
Is anyone working on continual yearning?
English
20
6
119
6.8K
Arjun Karanam
Arjun Karanam@QuantumArjun·
An analogy we love using is the intelligence vs experience axes of LLM ability. The labs are great at pushing models further and further along the axis of raw IQ - there’s no need to state how remarkable model improvements have been over the past few years, An orthogonal axis though is experience - a model that actually knows the ins and outs of your company, and learns from experience. That’s what’s really valued in the economy. Terrence Tao is great, but Terrence Tao 30 years on learning my accounting flows? Sign me up. Maybe a cop out response to Dwarkesh’s video, but I think the solution to get here is all of the above: scaling context windows so that you can get to longer sessions with an agent, dreaming and simulating environments to overcome sample inefficiency during training, and leveraging advancements on top of OPSD to actually credit sign properly. Combine all that together with an interface where anyone can teach their models like they do a human, and you get a peak into what the future of continual learning looks like
Dwarkesh Patel@dwarkesh_sp

What does the next training paradigm look like? 0:00:00 – The big research bet the labs are making 0:02:12 – Grindability is just as important as verifiability 0:06:10 – Will RLVR alone generalize? 0:08:41 – Getting the learning back to the weights 0:15:22 – Dreaming 0:17:23 – What 2027 looks like Also on YouTube, pod feed, and Substack.

English
2
0
21
2.2K
Lightspeed
Lightspeed@lightspeedvp·
Real-time world models represent a fundamental shift in AI. @reactorworld is building the platform for real-time generative video infrastructure, supporting developers who need the tech for use across entertainment, physical AI, and robotics. Co-founders @taiuti and @_bschmidtchen joined us last week on The Investment Memo, hosted by Partners @buckymoore and @theamberyang, to talk about the era of world models. The conversation centered around the infrastructure Reactor is building, why real-time models are the edge right now, and current use cases for the product. Alberto and Bryce agreed that world models are shaping the way simulations are created, and that developers need a streamlined platform that can support their ideas. We believe Reactor is positioned to be at the frontier of research into real-time generative models. We look forward to seeing how these models apply across industries. Chapters 00:00 Introduction & Overview of Reactor 01:08 Meet the Hosts & Founders 02:18 The Origin Story: From 3D Assets to World Models 05:07 Real-Time Video Applications Across Industries 06:55 The Open Source World Model Explosion 07:23 Why Infrastructure Is the Opportunity 08:42 Parallels to Past Technology Waves 09:51 Bridging the Research-to-Production Gap 13:13 What Developers Are Building with World Models 16:41 Lessons from Luma AI 18:23 What Apple Vision Pro Taught Bryce About Real-Time Systems 20:48 Company Values & Team Culture 22:40 Series A: What the Capital Unlocks 24:13 Reactor's Five-Year Vision 26:09 Closing Remarks
English
6
5
41
144.6K
Arjun Karanam retweetledi
Harvey
Harvey@harvey·
We partnered with @trajectorylabs to post-train NVIDIA Nemotron 3 Ultra for legal. Here’s what we found: 1) Open-weight models can reach frontier legal performance. On our Legal Agent Benchmark (LAB), Nemotron 3 Ultra started at a 0% all-pass rate. After post-training, it reached 5.8%, placing it between Sonnet 4.6 at 4.2% and Opus 4.6 at 6.6%. 2) Post-training dramatically improves reliability. Before training, many held-out tasks missed enough rubric dimensions to land around ~70% pass rates. After training, those tasks shifted toward ~95% pass rates. 3) Open-weight performance comes at much lower cost. Post-trained Nemotron 3 Ultra reached a similar quality band to leading closed models while running at roughly 1/8th to 1/50th the per-token price of Sonnet 4.6 and Opus 4.6. Most importantly: we post-trained this model on the @trajectorylabs platform less than 24 hours after Nemotron 3 Ultra launched, using the same harness, data, and recipe we used for Nemotron 3 Super. More to come as we continue to experiment with open-weight legal agents. Read more on post-training with Trajectory below:
Harvey tweet mediaHarvey tweet mediaHarvey tweet mediaHarvey tweet media
Trajectory@trajectorylabs

1/ We post-trained @nvidia Nemotron 3 Ultra on @harvey Legal Agent Bench in under 24 hours. The result: an open model reaching the same band as leading closed models on legal work, at a fraction of the cost. The correlating story: when a new open model ships, Trajectory can turn it into a specialized agent almost immediately.

English
11
33
287
52.8K
Arjun Karanam
Arjun Karanam@QuantumArjun·
The past few days made one thing clear: you have to own how your intelligence grows. We're thrilled to partner with Harvey to build that platform that turns models and harnesses from a fixed constraint into something you control
Trajectory@trajectorylabs

1/ We post-trained @nvidia Nemotron 3 Ultra on @harvey Legal Agent Bench in under 24 hours. The result: an open model reaching the same band as leading closed models on legal work, at a fraction of the cost. The correlating story: when a new open model ships, Trajectory can turn it into a specialized agent almost immediately.

English
0
1
17
1.1K
Sam Gorman
Sam Gorman@gormankind·
NASA's brand from 1977 still feels like it's from the future they even mapped out every detail in a 60 page manual
Sam Gorman tweet mediaSam Gorman tweet mediaSam Gorman tweet mediaSam Gorman tweet media
English
12
107
1.6K
64.7K
Arjun Karanam retweetledi
Jerry Tworek
Jerry Tworek@MillionInt·
That feeling when you realize practitioners say “verifiable tasks” when they actually mean “easy tasks.”
English
11
15
264
18.7K
fang
fang@sofangtastic·
i built an infinite canvas where your agent builds & publishes live web apps in real time works real-time with any coding agent that can add an mcp (codex, claude code, cursor, etc) great timing to ship this the day after OpenAI announced Codex Sites :D demo:
English
17
1
46
91K
Arjun Karanam retweetledi
Trajectory
Trajectory@trajectorylabs·
🏹 5 Days of Trajectory. Day 4 - Why We’re Building Trajectory AI is the most capable software ever built. You correct it. You teach it what you want. However, the next session starts, and the learning is gone. This is deeply unnatural - nothing intelligent works this way. Today, we’re sharing the thesis behind Trajectory: - why continual learning is the next platform shift in AI - why the primitive governing that shift is the trajectory - our plan to move products from being shipped to being grown: first make the intelligence layer better, faster, and cheaper; then make it shapeable; finally, make it learn Read more below⬇️
Trajectory tweet media
English
6
6
68
63.5K
Arjun Karanam retweetledi
Trajectory
Trajectory@trajectorylabs·
We’re taking a quick break for the 5 days of Trajectory, but wanted to take this time to say that we’ve been named to @Redpoint’s 2026 Infrared 100 as one of the companies shaping the future of AI infrastructure. We're so grateful for the recognition so early in our journey, and want to congratulate the other awardees as well!
Trajectory tweet media
English
5
5
69
8.3K
Arjun Karanam
Arjun Karanam@QuantumArjun·
Building the platform for continual learning means doing both cutting edge research and working with customers to make the right interface - so grateful that we've assembled a team that can do both
Trajectory@trajectorylabs

🏹5 Days of Trajectory. Day 3 - An Open Source Training Stack for Continual Learning Building the platform for continual learning requires both partnering with pioneering AI companies, as we showed on Day 2 with Harvey, and working toward frontier research, which we are highlighting today. Continual learning means models that improve hourly from real production use. But with the size of frontier models, this becomes quite difficult. A Qwen-397b would need to spin up and tear down repeatedly across six GPU nodes, and that's valuable time gone. Our contribution is Continual LoRA (C-LoRA): many lightweight adapters running at once on one shared base model. Our insight centers on where the parallelism lives: instead of splitting one giant job across nodes, we load-balance many small jobs over a single base. The result: 2.81x experiment throughput over single-tenant training, with no regression on rewards. We built this together, with @anyscalecompute, @NovaSkyAI, and generous support from @GoogleCloud and @GoogleStartups. We've open-sourced on SkyRL as one of the first multi-LoRA, RL training platforms, so that every team can get to continual learning faster. We’re very excited to see what you build, please reach out!

English
0
1
16
1.7K
Jake Paul
Jake Paul@jakepaul·
Met @MichaelElabd when he started Trajectory, excited to see what the team cooks up to make AI get better the more you use it! So excited for the Trajectory team, you guys move incredibly fast and are already pushing the frontier
Ronak Malde@rronak_

Today, @MichaelElabd, @QuantumArjun, and I are excited to announce Trajectory. We are a research lab and product company building the platform for Continual Learning. Our platform unlocks the signal already sitting in product usage, so companies can continuously post-train large-scale agentic models that outperform the frontier. @trajectorylabs We’ve raised $15M from @Conviction, @BessemerVP, @radicalvcfund, @jeffdean, @drfeifei and more. We’re partnering with some of the best AI-native companies: @ClayRunHQ @Harvey, @DecagonAI, @mercor_ai, @RogoAI to power their agentic systems, some of which we are already in production with. We’ve brought together a world class research team from DeepMind, OpenAI, Apple, Meta Superintelligence, Amazon AGI, Scale AI, and an elite product team from Stripe and Figma. AI will never again start on day one. Every correction, every retry, every edit will make products smarter. This is Continual Learning.

English
26
16
187
96.5K
Arjun Karanam
Arjun Karanam@QuantumArjun·
I first met @rronak_ and @MichaelElabd when we were all freshman at Stanford. Today, 8 years later, we’re announcing that we’ve started Trajectory, a research lab and product company building the platform for continual learning. I believe that Continual Learning demands a fundamentally new interface for how we build products. That's a research challenge and a product challenge in equal measure, so we've assembled a team to meet both: researchers from DeepMind, OpenAI, Apple, Meta Superintelligence, Amazon AGI, and Scale AI, and product talent from Stripe and Figma. We’re also partnering with the best AI native companies @Clay, @Harvey, @Decagon, @Mercor, and @RogoAI to power their agentic experiences, and push the boundaries of what agents look like in the real world. Please reach out if you’re excited to build with us!
Ronak Malde@rronak_

Today, @MichaelElabd, @QuantumArjun, and I are excited to announce Trajectory. We are a research lab and product company building the platform for Continual Learning. Our platform unlocks the signal already sitting in product usage, so companies can continuously post-train large-scale agentic models that outperform the frontier. @trajectorylabs We’ve raised $15M from @Conviction, @BessemerVP, @radicalvcfund, @jeffdean, @drfeifei and more. We’re partnering with some of the best AI-native companies: @ClayRunHQ @Harvey, @DecagonAI, @mercor_ai, @RogoAI to power their agentic systems, some of which we are already in production with. We’ve brought together a world class research team from DeepMind, OpenAI, Apple, Meta Superintelligence, Amazon AGI, Scale AI, and an elite product team from Stripe and Figma. AI will never again start on day one. Every correction, every retry, every edit will make products smarter. This is Continual Learning.

English
26
3
85
22.9K
Arjun Karanam retweetledi
Trajectory
Trajectory@trajectorylabs·
Welcome to Day 2. Yesterday, we showed the broader work we're doing with the pioneers of continual learning. Today we'd like to deep dive on one: how we post-trained an open model for legal work, in partnership with @Harvey. We've built a platform where production data is the moat. Every correction, retry, and edit becomes signal you can post-train on, and the models are plug and play: customer's can drop in their model of choice, and improve from there. Fields like legal and finance make those demands absolute, with hard security, sovereignty, and provenance requirements. That's why we post-trained @nvidia 's open-weight Nemotron 3 Super, on Harvey's LAB benchmark. The results, in just hours: post-trained Nemotron 3 Super approaches the closed frontier, matches GPT 5.5, lifts rubric-pass criteria +25%, all while beating the performance-vs-cost frontier. That's the power of our platform. And this is just a glimpse towards what the future of intelligence will look like: continual learning, where products get smarter every time they're used. Thanks to @nikogrupen, @gabepereyra, @ItsJulioPereyra, and the whole Harvey team for their collaboration on this. Much more to come soon on continually learning legal agents
Trajectory tweet media
English
12
7
112
66.7K

Keşfet

@code_star @lightspeedvp @reactorworld @taiuti @_bschmidtchen @buckymoore @theamberyang @trajectorylabs