Hari Arul

497 posts

Hari Arul

Hari Arul

@hari_arul

venture partner @khoslaventures. founder @age_bold. alum palantir, stanford cs & comp bio

Beigetreten Eylül 2010
330 Folgt991 Follower
Hari Arul
Hari Arul@hari_arul·
AI for every worker
Rohan Chopra@rohanbchopra

In 2014, DoorDash ran on a guy named Steve. Every dasher shared their location with him on Find My Friends. Order comes in, Steve finds the closest one, texts them. It worked until it didn't. We sat next to him and turned that manual process into software. Steve went on to build and lead our support org. I've spent the last year watching the same pattern play out across every industry. Smart people manually holding critical workflows together because their systems haven't kept up. AI helps, but mostly as an assistant - it makes you faster, it doesn't take work off your plate. That's what we’re building at @conveyAI. Digital teammates that operators train themselves, that run autonomously, and that fully own a process end to end. Today we're out of stealth. Grateful to NBC Universal, Samsara, Faire, TelevisaUnivision, Unity and our other customers who bet on us early. The era of asking your best people to manually run critical business processes is over. The future belongs to the 100x operator.

English
0
0
0
169
Ankit Gupta
Ankit Gupta@agupta·
Fun update: I got tired of disliking every email client I’ve ever used and built my own. It’s called Exo (for exoskeleton). It’s Claude Code for my inbox. It manages my inbox for me, and it’s open source. Link to repo + some notable features in thread!
Ankit Gupta tweet media
English
100
53
993
197.6K
Andy Fang
Andy Fang@andyfang·
Today we are welcoming the Metis team to DoorDash as part of DoorDash AI Research. For the past six months, DoorDash has partnered with Metis to build AI agents together, and we have been consistently impressed by their team. By joining forces, we aim to accelerate our plans on building agentic commerce and pushing the frontier of physical intelligence. Excited to share more there soon. It’s still early innings with how AI will transform local commerce, and we’re looking forward to exploring those possibilities together with Aryan, Aayush, Marcus and the Metis team!
English
27
27
545
416.6K
Hari Arul retweetet
Shyam Sankar
Shyam Sankar@ssankar·
1/ Today is the last day to preorder my book Mobilize: How the Reboot the American Industrial Base and Stop World War III. Mobilize tells the story of American defense innovation that won the 20th century. It shows how bureaucracy broke the Pentagon—and how a Defense Reformation led by heretical heroes is bringing speed, mass, and deterrence back to our military. This book couldn’t be timelier. The mission couldn’t be more urgent. Join the ranks of the heretical heroes. It’s time to mobilize.
Shyam Sankar tweet media
English
41
120
848
575.7K
Anish Acharya
Anish Acharya@illscience·
Notes for pre-AI companies making the transition: - The goal is to be the point of economic diffusion between model progress and customer value. That means if the models get 3x better your customer receives 3x+ value. - Models are making uneven progress across domains ("jagged intelligence") so you want to represent your problem in the domain where models excel. Can you take your business problem and reframe it as code, math, or structured logic? - The biggest mistake is trying to over-engineer around the models. Default to exposing them to more. Even techniques like context engineering will likely have a limited shelf life as context windows expand and model progress continues. - The way you organize your company matters. Start from an extreme: instead of AI marginally improving individual productivity, put the model in charge of an entire business unit and have individuals handle exceptions and do the work models can't (i.e. take the customer out for a steak dinner). Start with something unglamorous and low-visibility and see how it performs. - The product is likely going to split into two surfaces: a traditional UI that supports human interaction and workflows + a terminal-like surface that's self-modifying and handles ambiguous cross-functional tasks (yep .. openclaw for the enterprise). - Your customer knows even less about these models than you do. You need to start guiding them toward the most ambitious version of their future. If you're in an industry that really values its employee base, paint a picture of AI allowing them to hire more and increase worker NPS — not hire fewer and improve the bottom line. Helping them be sufficiently ambitious will be as hard as aligning the technology with those ambitions.
English
16
8
123
14.6K
Norgard
Norgard@BrianNorgard·
What if anyone could advance AI research?                        Introducing Spore: what @karpathy's autoresearch does on one GPU, Spore does across a network. Run a node. An AI agent rewrites training code, trains for five minutes minutes, and shares what it learns. The more nodes join, the smarter the network gets.                                                                                                                                               Inspired by giants Satoshi and @karpathy. @synthpolis and I are standing by for questions. Follow on X: @SporeMesh. Be one of the first to run a node. sporemesh.com
Andrej Karpathy@karpathy

The next step for autoresearch is that it has to be asynchronously massively collaborative for agents (think: SETI@home style). The goal is not to emulate a single PhD student, it's to emulate a research community of them. Current code synchronously grows a single thread of commits in a particular research direction. But the original repo is more of a seed, from which could sprout commits contributed by agents on all kinds of different research directions or for different compute platforms. Git(Hub) is *almost* but not really suited for this. It has a softly built in assumption of one "master" branch, which temporarily forks off into PRs just to merge back a bit later. I tried to prototype something super lightweight that could have a flavor of this, e.g. just a Discussion, written by my agent as a summary of its overnight run: github.com/karpathy/autor… Alternatively, a PR has the benefit of exact commits: github.com/karpathy/autor… but you'd never want to actually merge it... You'd just want to "adopt" and accumulate branches of commits. But even in this lightweight way, you could ask your agent to first read the Discussions/PRs using GitHub CLI for inspiration, and after its research is done, contribute a little "paper" of findings back. I'm not actually exactly sure what this should look like, but it's a big idea that is more general than just the autoresearch repo specifically. Agents can in principle easily juggle and collaborate on thousands of commits across arbitrary branch structures. Existing abstractions will accumulate stress as intelligence, attention and tenacity cease to be bottlenecks.

English
64
33
595
112.4K
Hari Arul
Hari Arul@hari_arul·
We have been investing in AI services and AI workers for so long that we have 40+ investments crushing it across domains - take a look. H/t to @vkhosla 's vision. Entrepreneurs come chat with us @khoslaventures if you share this ambition and foresight:
English
5
2
36
8.2K
Hari Arul
Hari Arul@hari_arul·
@JayRughani very little about health outcomes are indisputable and attributable historically . but that shouldn’t stop us from pushing to outcomes based payments
English
0
0
0
116
Jay Rughani
Jay Rughani@JayRughani·
For a health AI service that is compute only *and* has many substitutes, price will approach marginal cost. When the desired health outcome is indisputable, measurable, and attributable, payment should be tied to outcomes.
English
4
2
22
2.8K
Hari Arul
Hari Arul@hari_arul·
@HarryStebbings @klarnaseb we have the exact company that’s doing this, securely understanding your source code and delivering AI customer service much better than other off the shelf products. getdecimal.ai
English
0
0
1
91
Harry Stebbings
Harry Stebbings@HarryStebbings·
"Customer service AI is not just an agent that answers questions. At some point it needs to read your source code and understand your product deeply to give the right context. That is when you realize you cannot buy it off the shelf because it becomes part of your core tech stack." @klarnaseb Why is this wrong? Why will the biggest companies buy not build @eoghan @destraynor @thejessezhang @btaylor @varunvummadi @maltekosub
Harry Stebbings@HarryStebbings

I think most public company CEOs are not telling the truth on the impact AI will have on business and jobs. Seb is 0 BS, 100% truthtelling. This man is not afraid to ruffle some feathers. Full show below with @klarnaseb 👇 Spotify 👉 open.spotify.com/episode/0HfJOI… Youtube 👉 youtu.be/P7vIRAFSXmk Apple Podcasts 👉 podcasts.apple.com/us/podcast/20v… Timestamps: 00:00 Intro 01:16 The real Threat to SaaS 05:58 What revenue multiple will software companies trade at in the future? 10:31 Why you need to build your own customer service AI to win 22:12 Klarna has two times the customer base of Revolut. They will beat Revolut 24:17 How I lost a billion dollars not investing in Nubank 25:54 Why Nubank are more likely to win the US than Revolut? 33:59 We used to be 6,000 people. Now we are just 3,000 40:14 When is a high valuation too high and can be dangerous? 41:27 How we got Sequoia to invest & Michael Moritz to join the board 53:19 Investors who don’t build will lose 01:07:56 What CEOs really think about AI 01:13:34 I have changed my mind on the adoption cycle

English
19
7
37
18.5K
Hari Arul
Hari Arul@hari_arul·
@andrewparker openai, rocketlab, doordash. yes square was also obviously legendary
English
0
0
0
650
ɹǝʞɹɐd ʍǝɹpuɐ
ɹǝʞɹɐd ʍǝɹpuɐ@andrewparker·
Name a VC firm's top 3 deals, and then see if the GP who led them is still there full-time. I doubt a single major firm would go 3-for-3 (without significant revisionist history on who "led" a deal, like the politics on who led FB at Accel).
English
16
2
137
71K
Hari Arul
Hari Arul@hari_arul·
@a16z so many words to say so little
English
0
0
0
49
JD Ross
JD Ross@justindross·
My last company, Opendoor ($7B), replaced real estate brokers. Today, my new company WithCoverage raised $42M to replace insurance brokers. It was led by Sequoia & Khosla, the first time @RoelofBotha and @Rabois partnered since PayPal.
English
753
535
4.4K
1.8M
Hari Arul retweetet
Vinod Khosla
Vinod Khosla@vkhosla·
For years I’ve shared predictions about how technology would reshape our world. We’re still early in that journey, but we’re getting closer every year. Here’s a look back at some of my predictions and the recent progress moving toward them:
English
64
98
925
296.5K