RT @JustinWaugh: Love this. I’ve been building toward the same idea from another angle with Shikumi: persistent processes on the GPU, each…
This quoted post is unavailable.
Hadley Harris retweeted
Tiger Woods Unimpressed With Roadster As He's Flipped Dozens Of Vehicles buff.ly/dUGarJ0
Hadley Harris retweeted
Introducing Hugo: the first human companion robot.
At Boston Dynamics, I was Chief Strategy Officer, bringing robots into the real world.
My co-founder, Farbod Farshidian, led research at the RAI Institute.
Together, 15+ years teaching robots to run, jump, and work.
So we asked: what if robots could connect with people?
Hugo runs on the first Vision-Language-Expression (VLE) model.
He links what he sees and hears to movement, expression, and gesture.
He notices you, reads the room, and reacts in real time.
And he's only the first.
Dynamic Creatures is @BostonDynamics' official entertainment and hospitality partner.
We're building a cast of characters with film studios, resorts, and theme parks.
dynamiccreatures.com
Hadley Harris retweeted
We built this and will be giving agents access soon. We went through a lot of effort to have a clean product catalog, top tier name brand suppliers that can fulfill within 24 hours. LFG 🔥🔥🔥
The future is a merchant product that makes one’s inventory available via MCP / headlessly and accepts agent payments.
I suspect shipping can be done thru fedex/ups/usps direct from merchant at similar cost when you strip out the Amazon take rate (ie merchant can spend more if needed to ship quickly).
The moat isn’t what it is anymore. That is why they are blocking.
Hadley Harris retweeted
A year ago, Goodlabs was in one blood center, one day a week.
Today we're in 19 centers across six states, open 7 days a week, with more than 70,000 people on the platform.
Thank you to our donors and partners for making this possible.
Hadley Harris retweeted
Last night @human_ventures, @iconnections_io, and @EniacVC co-hosted 80+ family offices overlooking Union Square NYC, and it reminded me how much this kind of room is missing.
For years, we've poured energy into building founder community. Dinners, podcasts, intros, late-night texts. We know how much founders gain from being connected to each other.
But LPs? They're often doing this alone. Family offices are being asked to make sense of AI, shifting fund dynamics, and a very different venture landscape, usually without a real peer network to compare notes with. LPs need connectivity to each other now more than ever.
So we built the room. The conversations were candid, generous, and long overdue.
Huge props to @HeatherHartnett at Human Ventures and Kaitlin Malin at iConnections for bringing everyone together. 🙏🏾
More of these to come.
Hadley Harris retweeted
incredible things happening at the Actor x Pi hackathon this past weekend.
50 competitors, a collection of life sized and mini embodiments, and 36 hours to make them autonomous.
Needless to say the (autonomous) future is very bright!
some of my fav projects ⬇️
Hadley Harris retweeted
Just do things; @mikepat711 showing us the way
current AI state of affairs by @sundeep
"you can now build a startup with grokbot in your tesla... on the way to a startup event"
Hadley Harris retweeted
I agree with Dario that we need to pace the frontier. This has been a primary topic of discussions we've had at OpenAI in recent weeks.
Committing to having independent evaluators with employee-like access is a great idea, and we will do the same. We'll have more to share soon.
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Hadley Harris retweeted
We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so.
Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training.
You can read the full post here: darioamodei.com/post/we-must…
Hadley Harris retweeted
Today we're launching OM2.
Your AI re-reads your entire company from scratch every time you ask it something.
It’s why more than 50% of your token bill isn't in the answer, but in the search for it.
OM2 gives your AI a permanent memory of your company. 🧠
Hadley Harris retweeted
"Multi-agent systems" has officially entered the healthcare buzzword hall of fame, right up there with "value-based care" and "interoperability."
Everyone's throwing agents at everything and calling it innovation. But every so often, a company actually shows their work. And that's exactly what Predoc (@predoc_ai) did with a new report on how they built the multi-agent architecture behind their medical records retrieval and curation engine.
The full report (link below) is a useful blueprint for any healthcare exec trying to figure out what "building with AI" should actually look like in practice or where AI can have the highest impact.
4 things worth digging into:
1. Bespoke work is the opportunity, not the obstacle. Predoc makes the case that the most valuable automation opportunities aren't the clean, standardized tasks. Those get commoditized fast. It's the messy, facility-specific, exception-riddled workflows that are actually defensible. I think that's right, and it's a useful gut-check for any exec evaluating an AI vendor's ROI or worth buying.
2. The dataset is your moat. Predoc built its system on 300K-400K provider-research tasks, nearly 3 years of transcribed retrieval calls, and millions of reviewed record pages. The foundation models are swappable. That accumulated, structured "tribal knowledge" is not.
3. Start from first principles. Break the workflow down into its simplest parts. Bound each job, structure the handoff, escalate the exception. Predoc lays out how they gave each agent a job (research, voice, indexation, extraction, curation) and a structured output the next agent can act on immediately. When something doesn't fit, the agent escalates to a human, and that resolution gets fed back into the system.
4. The numbers back it up. I was pretty intrigued by some of the results in this piece: A 2-week-plus turnaround compressed to a median of 3 business days. Provider-research time down 70%. First-pass retrieval success up nearly 50%. 94.6% of pages indexed without human intervention.
The bigger theme I keep coming back to: this is a case study in systems of intelligence sitting on top of disorganized, disparate systems of record. Predoc's real output isn't "faster fax retrieval." It's a normalized, longitudinal clinical data layer that other applications can actually query.
Big thanks to brand partner Predoc for sitting down with me and showing their work on this one.
predoc.ai/multi-agent-system…
Hadley Harris retweeted
Amazing @truefoundry 's Open Source Harness- TrueForge is trending #1 on twitter!
Hadley Harris retweeted
Large language models are bad teachers, despite hundreds of millions of people using them to learn every day. Here are a bunch of reasons why:
1. They overwhelm you with huge blocks of text
2. They put the burden on you to decide what to do next
3. They veer off course without trying to keep you on track
4. They don't explicitly build and stick to a learning path like a teacher does
5. They force everything to happen inside of their experience
6. They speak robotically and pedantically
7. They don't easily let you create and manage helpful learning resources
8. They don't quantify your progress
9. They often attempt to create visuals via text, emoji, or ascii art circa 1999
10. They obfuscate how they know what they know
Oboe on the other hand:
1. Breaks everything down into clear, digestible chunks
2. Drives the conversation forward while allowing you to interject or ask questions
3. Has an opinion, like a good teacher does, of where to lead you
4. Creates a learning path, makes sure you see it, and evolves it as needed
5. Allows you to share and consume what you create in Oboe elsewhere, including podcasts
6. Speaks to you like a friendly, helpful tutor
7. Lets you create tons of artifacts, from diagrams to worksheets, from quizzes to songs
8. Helps you understand how much you're progressing across all your chats
9. Creates beautiful, modern, clear visuals only when it would help the learning process
10. Shares exactly what its sources are and how they tie into the material.
Check out our newest version:
Hadley Harris retweeted
Proud of portfolio company @LunarOutpostInc’s partnership with @nvidia to put GPU’s on the Moon!
Hadley Harris retweeted
Hermes from @NousResearch and @Teknium can now work through sites that are difficult to navigate.
The Steel browser plugin runs Hermes' tools in the cloud, with live viewing and page extraction.
It's Hermes browsing, turned all the way up.
See how to set it up ↓
Hadley Harris retweeted
We're throwing a Summer Interns Meet the C-Suite event on Aug 6th with @InspiredCap here in NYC! If you're an exec, this is your chance to meet AI-native interns who could be shaping your company's next decade. If you're an intern, this is your chance to meet execs, learn about what's keeping them up at night, and show off what you've built this summer🔥 Grab a spot 👇