Serving bold builders of the future. Not an offer to buy/sell securities or investment advisory services. Disclosures: https://xitter.cf/t.co/NqUCVhdnhA
Global
Joined July 2009
- Tweets9.9K
- Following2K
- Followers133K
- Likes8.7K
Pinned Tweet
Today we're announcing over $9B in capital for a new family of funds — our largest raise in 25 years.
AI is rewriting the rules. We're backing the founders building what comes next.
SpaceX wants to be your phone carrier.
The FCC approved SpaceX’s planned 15,000-satellite direct-to-device constellation. The company also agreed to acquire nationwide low-band spectrum, subject to FCC approval.
“I think this is really their effort to own the customer,” says @clairejyz on Lightwork. “Because right now they do have Starlink Mobile, but that is intermediated through T-Satellite and T-Mobile.”
@Machiz sees the potential to bundle mobile service with home internet. “You have your SpaceX mobile with your SpaceX Wi-Fi. And then all of a sudden, do you need some of these other things that you’re already paying for? Maybe not.”
But he cautions that building the service will likely take time. “You need to buy a spectrum and then you need to build an actual credible network.”
Check out the full Lightwork episode!
If labs set their own benchmarks, who defines the frontier?
We are seeing the top models sit within ~5 points of each other as labs often compete on specific benchmarks, how they sell, and price.
This week on Lightwork:
→ Gemini Argon vs. Muse Wonder: Optimizing for different workflows. Argon leans into legal, business, and finance. Wonder bets on long-horizon agent memory.
→ Apple HomeView: Privacy, personalization, and ambient intelligence could give Apple an opening in the AI-powered home.
→ SpaceX received FCC approval to launch 15,000 satellites in a push to becoming a major mobile carrier in the U.S.
0:00 Welcome to Lightwork
1:04 Meta Muse Wonder vs. Google Gemini 4
5:00 On AI Agent Memory and Focus
6:23 Gemini’s Advanced Reasoning and Task Automation
8:37 Apple HomeView: Siri Comes Home
11:33 AI Privacy and Smart Homes
16:52 SpaceX’s Plans To Be a Carrier
22:12 Updates from the Lightspeed Universe
@clairejyz @Machiz
AI is knocking on your front door.
Apple unveiled HomeView, a smart display that recognizes household members and brings Siri AI into the home, alongside an updated Apple TV 4K and HomePod mini 2.
On Lightwork, @clairejyz describes trading privacy for more capable agents on her phone. At home, it’s a different story.
“I don’t know if I’m ready to let Instinct or Claude into my home… But am I ready for it to have 24/7 ambient access into my conversations at home and pick up on things?”
She sees Apple’s emphasis on privacy as a potential advantage in earning that invitation.
@Machiz calls HomeView “a wedge into a very old market.” He sees an opening for Apple as other AI companies focus on apps and give the home experience less attention.
Check out the full Lightwork episode below!
If labs set their own benchmarks, who defines the frontier?
We are seeing the top models sit within ~5 points of each other as labs often compete on specific benchmarks, how they sell, and price.
This week on Lightwork:
→ Gemini Argon vs. Muse Wonder: Optimizing for different workflows. Argon leans into legal, business, and finance. Wonder bets on long-horizon agent memory.
→ Apple HomeView: Privacy, personalization, and ambient intelligence could give Apple an opening in the AI-powered home.
→ SpaceX received FCC approval to launch 15,000 satellites in a push to becoming a major mobile carrier in the U.S.
0:00 Welcome to Lightwork
1:04 Meta Muse Wonder vs. Google Gemini 4
5:00 On AI Agent Memory and Focus
6:23 Gemini’s Advanced Reasoning and Task Automation
8:37 Apple HomeView: Siri Comes Home
11:33 AI Privacy and Smart Homes
16:52 SpaceX’s Plans To Be a Carrier
22:12 Updates from the Lightspeed Universe
@clairejyz @Machiz
Engineers are pushing to build and ship hardware faster.
@Saronic’s co-founder and CTO, Vibhav Altekar, is sitting down with Lightspeed Partner @guruchahal to discuss what that takes in practice.
They’re welcoming engineers to the fireside chat and Q&A, “Ship Happens,” to hear what the team has learned building autonomous maritime systems and bring their own questions.
Join us this Thursday, Oct 22, in person at Saronic’s headquarters in Austin or virtually.
Sign up below to save your spot!
Curious what it takes to build at Saronic? Our CTO Vib Altekar will be joined by Lightspeed’s Guru Chahal to talk about the hard lessons of building hardware fast. Save your spot here: airtable.com/appIs0UG7L4Oz9S…
Lightspeed retweeted
A huge moment for the next generation of aviation automation. A $1B+ valuation plus two new offices to go after safer and more accessible flight systems.
Congrats @moof2k, Juerg, and Team @reliable_robot! We @lightspeedvp are proud to be your partners.
NEWS 📣 Reliable has raised new capital and officially surpassed $1B in valuation.
📍We’re expanding: new offices in Seattle, WA & San Diego, CA.
We're moving fast to build certified autonomous flight systems that make aviation safer and more accessible
reliable.co/pr/reliable-robo…
Lightspeed retweeted
“AI is going to take the old business models of ads and subscriptions and just annihilate them. Something will replace it, and I think there’s a good likelihood it’s going to be built on stablecoins.” - @eastdakota
Was incredibly prescient of Matthew Prince (@cloudflare) with @benmilne (@bralexyz) at @averystableconf in March of this year. Can't wait to tell you what @ay_o @kerrykellogg and I are working on for next year!
Booking a flight doesn’t require superintelligence.
Meta and Google have introduced models they describe as frontier, highlighting different strengths.
Meta’s Muse Wonder emphasizes memory and sustained task completion, while Google’s Gemini 4 Argon combines visual understanding, coding, and reasoning.
On Lightwork, @clairejyz questions how those frontier claims are judged. “Everyone’s a frontier when they’re benchmarking against whatever they’re picking and choosing.”
“It does seem like everybody’s benchmaxxing,” @Machiz adds.
For everyday users, the practical test is whether an assistant can remember what they need and finish the task.
“Can you help me book a flight? Can you help me clean up my emails? Those don’t require superintelligence. It requires focus and memory,” says Claire.
Tune in to the full episode of Lightwork!
If labs set their own benchmarks, who defines the frontier?
We are seeing the top models sit within ~5 points of each other as labs often compete on specific benchmarks, how they sell, and price.
This week on Lightwork:
→ Gemini Argon vs. Muse Wonder: Optimizing for different workflows. Argon leans into legal, business, and finance. Wonder bets on long-horizon agent memory.
→ Apple HomeView: Privacy, personalization, and ambient intelligence could give Apple an opening in the AI-powered home.
→ SpaceX received FCC approval to launch 15,000 satellites in a push to becoming a major mobile carrier in the U.S.
0:00 Welcome to Lightwork
1:04 Meta Muse Wonder vs. Google Gemini 4
5:00 On AI Agent Memory and Focus
6:23 Gemini’s Advanced Reasoning and Task Automation
8:37 Apple HomeView: Siri Comes Home
11:33 AI Privacy and Smart Homes
16:52 SpaceX’s Plans To Be a Carrier
22:12 Updates from the Lightspeed Universe
@clairejyz @Machiz
Lightspeed retweeted
1/ Today we're launching @different_ai , my third company, with $48M in seed funding led by @htaneja at @generalcatalyst, with participation from @ravi_lsvp at @lightspeedvp.
Different is an AI marketing system that runs demand generation end-to-end for lean B2B marketing teams.
Insurance was built to help millions of businesses take risks, but sometimes it can hold growth ambitions back.
Today, @rushilgoel launches @mandlhq to rebuild the full engine of commercial insurance with AI.
Mandl brings together three AI-native companies: insurance pricing (Nirvana), claims administration (Vega), and operations (Cooper).
We led @NirvanaTechInc’s seed round six years ago to help modernize a trillion-dollar industry mired in legacy processes. Mandl is the next step in that mission to price risk for what it really is, keep the promise fast, and get businesses back to their ambitions.
Congratulations to Rushil and the entire Mandl team on this exciting next chapter!
CC @ravirajjain, @AggShaurya
👋 Hello world! We're Mandl, a new AI company rebuilding the engine of commercial insurance from the inside out.
Insurance has always been the quiet force that lets ambition outrun risk: global trade, industrialization, and our modern economy were built on a foundation of innovative risk analysis and promises kept (the modern fire station got its start because of advancements in urban insurance!).
But as any hard-working business owner can attest, the system has now become just another risk that ambition has to outrun: overpriced and generic policies, delayed or denied claims. Even everyday things like pulling a COI still take forever.
At Mandl, we are taking on this entire system, part by part:
→ Underwriting: To advance our understanding of risk, we built @NirvanaTechInc. Now just a few years in, we're leveraging more than 50 billion miles of driving data, all matched to real claims outcomes and the most intelligent models in the industry. The business has grown to more than $300 million in run-rate premium with top-tier loss ratio performance.
🤝 Honesty, it turns out, is the best policy.
→ Claims: A claim is the moment you learn whether your insurance company keeps its promise to get you back to business. In less than two years, Vega has already made claims resolution 4x faster than the industry standard. We're proud of that, but we are not done advancing commercial resilience.
⚡️ Our mission is to resolve a claim in minutes.
→ Operations: To unleash insurance professionals so they can do their best work, we built Cooper, an AI operations platform for the commercial insurance office.
💰 It's early days, but we've already helped customers process more than $1 billion in premiums.
While we're only just getting started, we have ambitions to go deeper where we already operate and expand into more of the industry, investing wherever AI can make commercial insurance work the way it should.
Thanks to our team, customers, and partners for building with us, and to our world-class investors @generalcatalyst, @lightspeedvp, and Valor Equity Partners for backing us along the way.
You can read more from our CEO & cofounder @rushilgoel: mandl.ai/news/introducing-ma…
Coding agents have changed how software is written.
Today, @doubleAI introduces Code Execution Models, a new category of AI that is built for how software runs.
“The first AI that lives inside running code.”
CEMs place intelligence between code and silicon. The first CEM, doubleAI's Maxwell, continuously inspects and rewrites code inside the runtime, fitting it to the hardware, data and usage it encounters.
This makes code dynamic. CEMs get more out of every piece of hardware, so the same CPU or GPU can do more work, or the same work can be done with less hardware.
Congratulations to co-founders @AmnonShashua, @shai_s_shwartz, @laginimaineb, @YoavLevine, Dr. Noam Weis, @or_sharir, and the entire doubleAI team on the announcement!
CC: @YoniCheifetz
Beam’s release puts @reflection_ai’s research in more builders’ hands.
Congratulations to @MishaLaskin, @real_ioannis and the entire team. We’re incredibly proud to support!
Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners.
We're releasing:
- Weights, tokenizer, and reference implementation
- BF16, FP8 and NVFP4 weights and a DFlash2 drafter
- The open source harness we built and trained Beam on: Mirror CLI
Weights and model card on Hugging Face: huggingface.co/reflection
Lightspeed retweeted
Neko CEO Hjalmar Nilsonne was licking his wounds after "crash landing" a climate startup. Then he got a fateful DM from Spotify's Daniel Ek
Together, @HjalmarNilsonne and @eldsjal co-founded @Neko, a $7B startup that calls itself "the most valuable hour you can spend on your health."
After years operating in Europe, Neko recently launched in NYC, where it's quickly built a waitlist of 50,000 people, and started work on 3 more locations. Miami, San Francisco and Washington, D.C. are next.
“People sometimes think that there’s some machine, and you go in, and you get a kind of paper receipt telling you if you’re gonna die or not,” Hjalmar tells me. “In reality, it’s the exact opposite.”
On The Upstarts Podcast, Nilsonne talks about the fateful Twitter DM from Ek that got Neko going; why the doctor visit hasn’t changed much in 100 years; and how to keep a startup team motivated while chasing a long-term “moonshot.”
Plus, he shares his Upstart Moment: a bet-the-company pivot driven by an internal memo he wrote, “Free Is Not Enough.”
A big thanks to this season's sponsors @Atlassian and @chargebee for supporting unfiltered founder interviews like this 🫡
CHAPTERS:
00:00 Introduction
01:48 Mapping Your Body In One Hour
03:54 Why The Doctor’s Visit Hasn’t Changed In 100 Years
07:31 Building Four Businesses In One: Devices, Hardware, Software, Clinics
09:22 Breaking Down A Neko Visit First-Hand
12:58 Why Half of Neko’s U.K. Members Never Had A Health Check
16:59 ‘Brutally Efficient With Your Time’: Why Every Second Counts
19:04 Helping Human Doctors, Not Replacing Them
22:40 Delivering Outcomes For the 92% Of Members Not At Risk
26:47 A Climate ‘Crash Landing’ And A DM From Daniel Ek
33:12 Motivating A Team To Chase A Moonshot
37:42 Neko’s Upstart Moment: The ‘Free Is Not Enough’ Memo And Pivot
41:30 Pricing Neko Accessibly For The 99.9%
45:44 What The Tesla Roadster Can Teach Founders About Change
48:33 Space-Age Scan, Old-School Advice: Cut The Coffee
Thanks for watching!
If labs set their own benchmarks, who defines the frontier?
We are seeing the top models sit within ~5 points of each other as labs often compete on specific benchmarks, how they sell, and price.
This week on Lightwork:
→ Gemini Argon vs. Muse Wonder: Optimizing for different workflows. Argon leans into legal, business, and finance. Wonder bets on long-horizon agent memory.
→ Apple HomeView: Privacy, personalization, and ambient intelligence could give Apple an opening in the AI-powered home.
→ SpaceX received FCC approval to launch 15,000 satellites in a push to becoming a major mobile carrier in the U.S.
0:00 Welcome to Lightwork
1:04 Meta Muse Wonder vs. Google Gemini 4
5:00 On AI Agent Memory and Focus
6:23 Gemini’s Advanced Reasoning and Task Automation
8:37 Apple HomeView: Siri Comes Home
11:33 AI Privacy and Smart Homes
16:52 SpaceX’s Plans To Be a Carrier
22:12 Updates from the Lightspeed Universe
@clairejyz @Machiz
Going public takes more than strong demand. You need to know when to set sail.
Oura postponed its Nasdaq IPO hours before pricing, despite reported investor interest.
“It’s a bit of a mystery. I’m scratching my head a little bit myself,” says @AggShaurya.
On Lightwork, he and @Machiz discuss what it takes for a company to go public, from valuation and pricing to market timing.
Josh emphasizes how much market conditions matter: “It’s almost like going out for a day on the water. You need to be sure that all of the things are looking really perfect before you go out and sail.”
Tune in to Lightwork for the full conversation.
Two new open-weight models in two days.
Lightspeed portfolio companies Reflection and Mistral unveiled Beam and Large 4, all striving to bring greater control, customization, and efficiency to enterprise AI.
More on Lightwork from @Machiz and guest co-host, Partner @AggShaurya:
→ DoorDash builds its own delivery drone and AI ordering agent.
→ Instinct brings agents into group chats as Meta and Sierra develop new rules for agent access.
→ Oura postpones its IPO despite reported strong demand.
00:00 Open Models, Drone Delivery, and AI Agents
01:08 Mistral's Le Chonk Meme and Naming Story
02:33 What Open-Weight Models Mean for the AI Race
03:56 When to Consider Open Models Over Frontier AI
05:30 Why Open Models Compete on Cost Efficiency
06:35 Mixture of Experts Explained
07:32 Is Reflection the DeepSeek of the West?
07:58 Why DoorDash Is Building Its Own Delivery Drones
09:38 DoorDash's Competitive Edge
12:28 DoorDash's Moat Against AI Agents
14:34 Personal AI Agents in Group Chats
16:08 How AI Agents Could Change Consumer Spending
17:26 Meta, Sierra, and the Future of Personal AI Agents
19:56 Oura Postpones Its IPO
22:13 Lightspeed Universe
David Silver got his first computer at age seven. He's been working on the same question ever since:
"What is intelligence, and what's a mind, and how can you build one?"
That question led him to a university friendship with @demishassabis, co-founding a video games studio with him, and eventually DeepMind, where David led the teams behind AlphaGo and AlphaZero.
Now it's led him to @IneffableLabs.
In a new Lunch with the @FT, @madhumita29 sat down with David to discuss AI's next chapter: an "era of experience," in which systems learn on their own rather than relying on human data.
"There's a certain amount of human data that's out there on the internet — you can think of it like a fossil fuel. And once you've trained on it all, the fossil fuels are all gone," said David.
We're proud to partner with David and the team in their pursuit of a system that learns from its own experience.
Check out the full story below.
Read more in The Financial Times: ft.com/content/462c303b-b96c…
