←
Electric Capital ⚡️537 posts
Electric Capital ⚡️ avatar✓

Electric Capital ⚡️ ✓

@ElectricCapital

We're hiring. Electric Capital is an early stage venture firm focused on cryptocurrencies, blockchain, fintech, and marketplaces.

◷ Joined May 2018▧ 5 Photos and videos
Avichal Garg - Electric ϟ Capital avatar
⟳ Electric Capital ⚡️ retweeted
Avichal Garg - Electric ϟ Capital@avichal·18h

Well-Behaved Founders Seldom Make History - Most people make a fundamental mistake when they think about founders. They assume founders are business people. Great founders are not business people. Founders are artists who happen to be good at business. Business people operate within the world as it exists today and work forward. They see the world as it is. Founders imagine a world that doesn't exist yet. They become obsessed with the world as it could be. Founders make the world inside their heads come to life. This is what artists do. And artists are often very strange people. Nobody expects van Gogh, Kurt Cobain, or Norm Macdonald to be balanced. We instinctively understand that the same traits that make someone capable of creating extraordinary things likely make ordinary life difficult. We understand artists are prone to addiction, mental health issues, and fringe beliefs. But somehow people expect founders to be different. They want founders to be visionary but not crazy. Confident but not egotistical. Risk-taking but not reckless. Unconventional in their ideas but conventional in their personal lives. Sometimes this works. I know founders of $10b+ companies who are genuinely happy, well-adjusted people with wonderful families. But I also know that if you were to sample the top 1,000 founders and compare to the general population, you would find a dramatically higher rate of strange beliefs, what most mainstream people would term mental health issues (though I disagree), and very alternative lifestyle choices. The greatest founders tend to be genuinely strange people, and I love them for it. For most outlier founders, creating and competing isn't really a choice. It's a compulsive behavior. It just happens to manifest through code and companies rather than on a canvas. And the qualities that make them great at disruption and creation can cause enormous problems elsewhere in their lives. This does not mean we should glorify or emulate their mistakes for the sake of it. Nor does it mean suffering is a prerequisite for genius. It means that the traits we value in extraordinary creators are not always separable from the traits that make them difficult people. Do we get the mobile revolution without Steve Jobs? Do we get to Mars without Musk? Do we AI that gives everyone a doctor in their pocket without Jensen? I don't think we do. And I don't think we should confuse being good at life with being good at creating technologies that change the world. Fundamentally, we can have art that moves us and technology that moves us forward. But if we want these benefits, we have to accept that many of the people best suited to this work are deeply flawed individuals. Loving their art does not mean we condone an artist's moral failings. If anything, understanding the artist as an imperfect, sometimes broken, soul should elicit compassion, love, and deep gratitude for the great benefits we receive as a society at their personal expense.

Rocket avatar
⟳ Electric Capital ⚡️ retweeted
Rocket@userocket_app·Oct 26

rocket public beta is live, now on monad. $10,000+ competition starts today. predict on testnet, win real prizes. more details below.

0:30
Kuru avatar
⟳ Electric Capital ⚡️ retweeted
Kuru@KuruExchange·Oct 19

Stocks. ETFs. Gold. All onchain. All on Monad. Introducing the new Kuru. Rebuilt from the ground up, with faster infrastructure, better execution, and a cleaner trading experience.

0:38
Quest avatar
⟳ Electric Capital ⚡️ retweeted
Quest@QuestHarness·Oct 13

Release 0.3.0 tagged! Major changes: - A new native docs artifact (Quest docs) that protects information movement based on visibility - Many new inference backends (@FireworksAI_HQ, @near_ai, @TrustedRouter) - Multiple accounts per user for connected services. - Haiku 5.5 and Argon 4 support for Vertex - Sanitization framework for model-created files. - New contributors! (Thank you @jperla and @altantutar) - Bug fixes, see full changelog below github.com/electric-capital/…

Ken Deeter (building @QuestHarness) avatar
⟳ Electric Capital ⚡️ retweeted
Ken Deeter (building @QuestHarness)@puntium·Oct 13

We now have a dedicated page for @QuestHarness! And new release tagged! This is the first release with external contributors. More inference backends. IMAP support. And a new agent-native document creation feature called Quest docs. 👇

Ken Deeter (building @QuestHarness) avatar
⟳ Electric Capital ⚡️ retweeted
Ken Deeter (building @QuestHarness)@puntium·Sep 28

While building Quest, I realized people often conflate two important facets of agentic permissions. One facet is the systemic one. "Can the agent send a Slack message?" "Can the agent write to this file?" We often think about these permissions in the context of a specific workflow or task. We build systems that say for workflow X, the agent can call function Y. We map the tools that an underlying system provides us into a permission layer for the AI. Which is great; it's part of the solution, but if you stop there you're missing half the story. Because the second facet is: "Even if the agent can write to this file, should it write what it's about to write?" Here's where it gets way more nuanced. The agent just used your CFO's access permissions to read everyone's salary. Can it use that information to write it to a file? If it does, who can see that file? Is there a violation of privacy, confidentiality, or company rules if it does? As you can see, just thinking about things at the "can it write to a file" level doesn't address a huge set of potential issues. And unlike these tool-level access controls, the rules of information flow within an organization are often fluid, poorly documented, and highly contextual. This is why in Quest, we went with a strict human-in-the-loop design for all scenarios where an AI can share information retrieved from your context with other people/agents. Someday, we hope to introduce mechanisms to allow AIs to help you make these decisions. If these gatekeeper AIs get good enough, then you might imagine some of this even gets automated. But for this to work there's a whole set of skills and context that needs to be built for every organization according to their own rules. And then there's the whole layer of evals and testing needed to measure resilience against prompt injection and alignment behavior. All before we can ever hope to turn this on for handling the most sensitive data. Until then, the agent can still do a lot of work for you, but I think you should have the final say.

Post attachment
Avichal Garg - Electric ϟ Capital avatar
⟳ Electric Capital ⚡️ retweeted
Avichal Garg - Electric ϟ Capital@avichal·Sep 23

We @ElectricCapital are open sourcing Quest - a harness for privacy-aware companies Run your own models, be thoughtful about which data is shared with model companies, & guardrails to make sure non-technical users don't accidentally leak sensitive data We use it every day 👇

Ken Deeter (building @QuestHarness) avatar
⟳ Electric Capital ⚡️ retweeted
Ken Deeter (building @QuestHarness)@puntium·Sep 23

Today we are open sourcing Quest, an AI harness that we built at @ElectricCapital, used by all of our teams every day. Deploying AI to a whole organization is not easy. We looked at many, many options. Ultimately, there were two reasons that led us to build our own harness: 🧵

ryoh 🕊️ avatar
⟳ Electric Capital ⚡️ retweeted
ryoh 🕊️@ryohhno·Sep 3

Since launch 4 months ago, we've been testing and iterating and building on what makes Legend the greatest trading platform in the world. We believe that we've now found it and are officially launching the product out of Alpha. - We've done ~$800M in trading volume - We're now onboarding ~1,000 signups a day - Our initial feature set is built and ready to be scaled - [redacted] coming soon And most importantly of all, we've found a group of users/friends who want to see us win and believe in what we're building. A place for pro traders and new traders alike to become Legend.

0:41
Monad avatar
⟳ Electric Capital ⚡️ retweeted
Monad@monad·Sep 1

Introducing the Metropolis global hackathon Six weeks to build the next generation of startups on Monad. 4 tracks. Over $250K in prizes for builders. Submissions end October 13th. monad.xyz/metropolis

0:35
V12 avatar
⟳ Electric Capital ⚡️ retweeted
V12@v12sec·Jul 30

luba yudasina avatar
⟳ Electric Capital ⚡️ retweeted
luba yudasina@LubaYudasina·Jul 23

Today I sit down with @avichal , co-founder and General Partner at @ElectricCapital . Avichal has built and sold companies, worked inside early Google and Facebook, served as a visiting partner at Y Combinator, and now backs founders across crypto and frontier tech. We get into: • Why founders are more like artists than operators • What actually compounds a person's trajectory • His "software eating money" thesis behind Electric Capital • Why honesty ends up being the real edge in venture • so much more! Timestamps 00:00:00 Trailer 00:00:44 Intro 00:01:00 Watches, luxury markets, and the AI parallel 00:12:48 Opportunity is a luxury 00:14:43 Founders as artists, not just operators 00:18:26 Authenticity and cutting through market noise 00:24:38 Betting on outliers as a life philosophy 00:37:58 Storytelling as a founder superpower 00:55:41 How Elad Gil helped launch Electric 01:07:55 Why crypto was the wedge for Electric 01:19:28 Why he prefers market corrections 01:31:00 The firms Electric was modelled after 01:36:21 Honesty as the core operating principle 01:42:15 Over-committing resources to founders 01:51:03 Dodging SBF and reading fake authenticity 02:01:40 Outro Link in comments!

2:01:54
Curtis Spencer avatar
⟳ Electric Capital ⚡️ retweeted
Curtis Spencer@curtis0x·Jul 16

Love Linus' pragmatism here: "AI is a tool, just like other tools we use." The data agrees: 1 in 12 Linux kernel commits merged last month disclosed AI assistance. A year ago that number was zero.

0:10