Sabitlenmiş 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.
English
Lightspeed
9.6K posts
Lightspeed
@lightspeedvp
Serving bold builders of the future. Not an offer to buy/sell securities or investment advisory services. Disclosures: https://t.co/NqUCVhdnhA
Global Katılım Temmuz 2009
2K Takip Edilen102K Takipçiler
Lightspeed retweetledi
More than 80% of US physicians are already leveraging AI. The question now is how consumer adoption of health AI is going to drive real patient outcomes, with companies like @Doctronic & @Neko helping lead the charge to deliver both access *and* accuracy.
I had a chance to discuss how AI is more deeply integrating across the patient journey alongside my partners Dr. @GalymImanbayev, @clairejyz, and @Machiz on @LightspeedVP’s new show Lightwork.
Check out the full video to hear more about our thesis below!
cc @BrentonFargnoli
Claire Zau@clairejyz
NEW EPISODE: @AnthropicAI looks inside AI's mind with the “J-Space,” the $500M fund to end the common cold, and @Amazon’s Project Leo challenges @Starlink @Lightspeedvp Partners @GalymImanbayev and @f_fatemi21 map the AI-native patient journey, and the companies we see innovating in the healthcare x AI space 🧬🧑💻🧪. Plus a deep dive into @Doctronic's AI doctor and @Neko's vertically integrated approach to preventive care. 0:00 Meta Muse Image Controversy 8:34 Anthropic's "J-Space": Can We Read An AI's Mind? 17:27 Silicon Valley Wants To End The Common Cold 23:30 Amazon Takes On Starlink 29:13 Mapping AI Across The Patient Journey 45:36 Doctronic 53:33 Neko Health 1:01:09 Startup Opportunities In The Patient Journey @Machiz
English
Do models have something like a subconscious?
@clairejyz on Anthropic’s research into J-Space and the hidden processing behind model behavior:
"For many years, people have been trying to understand how a model's brain works."
"Could a model have a divide the way humans do between accessible thoughts and all the unconscious, below-the-surface processing?"
"The J-Space is, in essence, a mental whiteboard."
"In the same way that humans have a mental whiteboard where a lot of thoughts are processed behind the scenes, models have a similar concept."
Claire Zau@clairejyz
NEW EPISODE: @AnthropicAI looks inside AI's mind with the “J-Space,” the $500M fund to end the common cold, and @Amazon’s Project Leo challenges @Starlink @Lightspeedvp Partners @GalymImanbayev and @f_fatemi21 map the AI-native patient journey, and the companies we see innovating in the healthcare x AI space 🧬🧑💻🧪. Plus a deep dive into @Doctronic's AI doctor and @Neko's vertically integrated approach to preventive care. 0:00 Meta Muse Image Controversy 8:34 Anthropic's "J-Space": Can We Read An AI's Mind? 17:27 Silicon Valley Wants To End The Common Cold 23:30 Amazon Takes On Starlink 29:13 Mapping AI Across The Patient Journey 45:36 Doctronic 53:33 Neko Health 1:01:09 Startup Opportunities In The Patient Journey @Machiz
English
Lightspeed retweetledi
HEALTHCARE x AI: THE PATIENT JOURNEY
@GalymImanbayev and @f_fatemi21 broke down all 7 stages, from "get covered" to "get paid," and companies building in each one
(1/9)👇
Claire Zau@clairejyz
NEW EPISODE: @AnthropicAI looks inside AI's mind with the “J-Space,” the $500M fund to end the common cold, and @Amazon’s Project Leo challenges @Starlink @Lightspeedvp Partners @GalymImanbayev and @f_fatemi21 map the AI-native patient journey, and the companies we see innovating in the healthcare x AI space 🧬🧑💻🧪. Plus a deep dive into @Doctronic's AI doctor and @Neko's vertically integrated approach to preventive care. 0:00 Meta Muse Image Controversy 8:34 Anthropic's "J-Space": Can We Read An AI's Mind? 17:27 Silicon Valley Wants To End The Common Cold 23:30 Amazon Takes On Starlink 29:13 Mapping AI Across The Patient Journey 45:36 Doctronic 53:33 Neko Health 1:01:09 Startup Opportunities In The Patient Journey @Machiz
English
YouTube
English
NEW: $500M is being poured to end respiratory diseases.
This week, @clairejyz and @Machiz unpack why some big names in tech are backing Intercept, a new initiative focused on making illnesses like the common cold, flu, and COVID far less common.
Lightspeed Partners @GalymImanbayev and @f_fatemi21 break down the AI care model, how they see healthcare going from being reactive to proactive, and companies like @Doctronic and @Neko are shaping consumer health.
They also dive into:
• How AI could change advertising with the first media generation models from Meta Superintelligence Labs
• Anthropic’s latest interpretability research and the race to understand what AI models are “thinking”
• Amazon Leo reaching a major milestone as a potential Starlink competitor
Full video link below.
Claire Zau@clairejyz
NEW EPISODE: @AnthropicAI looks inside AI's mind with the “J-Space,” the $500M fund to end the common cold, and @Amazon’s Project Leo challenges @Starlink @Lightspeedvp Partners @GalymImanbayev and @f_fatemi21 map the AI-native patient journey, and the companies we see innovating in the healthcare x AI space 🧬🧑💻🧪. Plus a deep dive into @Doctronic's AI doctor and @Neko's vertically integrated approach to preventive care. 0:00 Meta Muse Image Controversy 8:34 Anthropic's "J-Space": Can We Read An AI's Mind? 17:27 Silicon Valley Wants To End The Common Cold 23:30 Amazon Takes On Starlink 29:13 Mapping AI Across The Patient Journey 45:36 Doctronic 53:33 Neko Health 1:01:09 Startup Opportunities In The Patient Journey @Machiz
English
Europe's defense infrastructure is being rebuilt.
The New York Times @dealbook covered @HelsingAI’s drone factory in Germany.
When @torstenreil, Niklas Köhler, and Dr. Gundbert Scherf started the company in 2021, democracies were looking to future-proof their defenses.
Today, Helsing has assembled a strong management team that is embedding AI-enabled software in next-generation defense technology.
We believe Helsing is a leader in building intelligent autonomous systems and are excited to support their journey.
Read more via the link below.
@ravi_lsvp, @ravirajjain, @alex_schmitti
English
We’re proud to participate in @HelsingAI’s $1.8B Series E, bringing the company to a total valuation of about $18B.
As software develops further, so too must the technology that defends our infrastructure.
We first backed co-founders @torstenreil, Niklas Köhler, and Dr. Gundbert Scherf for their Series C in 2024. Since then, Helsing has tripled in valuation.
Helsing’s commitment to mission-ready, AI-native software for the defense industry has led to strong partnerships with hardware companies and governments across Europe.
We are very excited to continue partnering with the Helsing team. Congratulations on the raise!
CC: @ravi_lsvp, @ravirajjain, @alex_schmitti
English
Lightspeed retweetledi
The best thing happening in healthcare AI right now is how much the industry is choosing to build together.
Proud to co-host our first @AbridgeHQ hackathon with @AnthropicAI , @lightspeedvp , and @cerebral_valley . This is what positive-sum looks like.
cerebralvalley.ai/e/abridge-hack…
English
Arena started as “a side project of another project.”
In spring 2023, after Meta released Llama, Berkeley students spent a weekend fine-tuning it on shared GPT conversations. The result was Vicuna.
They put it on the web and believed it was better than base Llama. But then came the harder question: “How do you show that it’s better?”
Existing benchmarks were mostly “static tests.” They did not capture real interaction between humans and LLMs.
So they built what they called “battle mode”: ask anything, compare two randomized models, and vote.
“That’s how Arena was born.”
@istoica05 on Lightwork
Claire Zau@clairejyz
NEW EPISODE: @Meta reads your brain with Brain2Qwerty. @AnthropicAI releases Claude Science, and Co-founder and Chairman of @arena Ion Stoica joins us. @istoica05 is the Berkeley professor behind an absurd run of systems that basically run modern AI: @ApacheSpark (the standard for big data), Ray (what OpenAI used to train ChatGPT), and @vllm_project (the most-used LLM inference engine). He also co-founded @databricks and @anyscalecompute - and now chairs Arena, the internet's default scoreboard for AI models. 0:00 The AI memory squeeze 2:55 Brain2Qwerty: Meta Reads Your Mind 9:38 Anthropic Releases Claude Science 15:58 Why Every Lab Wants Pharma 20:52 How Chatbot Arena Was Born 26:58 How Does Arena Actually Rank AI Models? 31:11 Subjective vs Objective AI Quality 33:49 How Arena Evaluates AI Agents 37:54 How Does Arena Stay Neutral? 43:31 Can AI Become Unmeasurable? @Machiz @lightspeedvp
English
Lightspeed retweetledi
We raised $136M to kill Slack.
Introducing PromptQL: The first AI version of Slack.
Here’s how it works:
English
We believe great consumer hardware needs three kinds of genius in the room. Hardware, software, and design.
That’s why we’re excited to lead @Nirva’s $8M Seed Round.
Nirva seeks to design the world's most beautiful and insightful AI wearable.
Consumers are spending billions trying to optimize themselves. But AI wearables only get smart if you wear it.
We believe co-founders @realweilyu, @jasonqchen, and @HanyingHu have the kind of genius to make a wearable that people choose to wear everyday.
Wei and Jason’s experience building hardware and software products at Meta, combined with Hanying’s design DNA can bridge the gap that we believe consumer hardware demands.
Nirva understands how difficult it is to understand one’s own patterns and chronicle them. Their AI wearable was built to track users’ moods throughout the day automatically, mirroring their lives in the form of a journal.
The next iconic consumer device may not look like a gadget at all. Instead, it may look like a beautiful piece of jewelry.
We are incredibly excited to partner with the Nirva team. Congratulations on the launch.
CC: @LightspeedIndia, @PinnLawjindakul, Chris Halim, @kaluwi, and @kumarharsha2212
English
Learn more in Orca’s press release here: orcabio.com/orca-bios-treg…
English
.@OrcaBio just received FDA approval for their new immunotherapy drug, TREGZI.
This new precision-engineered cell therapy aims to help patients with blood cancer strengthen their immune systems.
Many patients with blood cancers are at risk of developing chronic graft-versus-host disease (GVHD) following a donation, and Tregzi is designed to use stem cells to help the immune system fight back.
Congratulations to founders Ivan Dimov, Dr. Nate Fernhoff, Jeroen Bekaert, and the entire Orca Bio team on this accomplishment!
We first partnered with the team when investing in their Seed round back in 2016. It has been incredible to see them grow and continue to work towards their goals to aid the treatment of cancer and autoimmune diseases.
CC: Partner @AggShaurya and Venture Partner Jonathan MacQuitty
Orca Bio@OrcaBio
We are proud to announce the U.S. FDA approval of our new precision-engineered cell therapy for allogeneic transplant in people with hematologic malignancies. This marks a significant milestone for Orca Bio, and most importantly, for the patients we serve. Learn more: bit.ly/4vGQOU0 #CellTherapy #BloodCancer
English
"A single new drug costs around $2.6 billion, takes 10-15 years, and up to 90% of candidates that enter trials fail."
@clairejyz on why the endgame for AI labs could be co-developing drugs with pharma:
"Whether it's Claude Science or GPT-Rosalind, they can tackle that front end. They can help with discovery, hit finding, lead optimization, predicting how it will play out in preclinical."
"If you can help with that part, then by the time you actually release a drug, you'll have a better hit rate. That's the flywheel they're going after."
From Lightwork
Claire Zau@clairejyz
NEW EPISODE: @Meta reads your brain with Brain2Qwerty. @AnthropicAI releases Claude Science, and Co-founder and Chairman of @arena Ion Stoica joins us. @istoica05 is the Berkeley professor behind an absurd run of systems that basically run modern AI: @ApacheSpark (the standard for big data), Ray (what OpenAI used to train ChatGPT), and @vllm_project (the most-used LLM inference engine). He also co-founded @databricks and @anyscalecompute - and now chairs Arena, the internet's default scoreboard for AI models. 0:00 The AI memory squeeze 2:55 Brain2Qwerty: Meta Reads Your Mind 9:38 Anthropic Releases Claude Science 15:58 Why Every Lab Wants Pharma 20:52 How Chatbot Arena Was Born 26:58 How Does Arena Actually Rank AI Models? 31:11 Subjective vs Objective AI Quality 33:49 How Arena Evaluates AI Agents 37:54 How Does Arena Stay Neutral? 43:31 Can AI Become Unmeasurable? @Machiz @lightspeedvp
English
Some say stronger models should make human evaluation less necessary. @istoica05 argues the opposite:
"Music is totally subjective. A smart LLM can maybe guess what I would like, but there is not only one answer. That comes down to human taste, and that will remain for a long while."
"Think about the most successful AI applications today. Coding: it's a developer in the loop. Customer support: it's satisfying a human and answering human questions."
"Despite the fact that these models are going to become more and more powerful, human signals are going to be very important, if not even more important."
From Lightwork
Claire Zau@clairejyz
NEW EPISODE: @Meta reads your brain with Brain2Qwerty. @AnthropicAI releases Claude Science, and Co-founder and Chairman of @arena Ion Stoica joins us. @istoica05 is the Berkeley professor behind an absurd run of systems that basically run modern AI: @ApacheSpark (the standard for big data), Ray (what OpenAI used to train ChatGPT), and @vllm_project (the most-used LLM inference engine). He also co-founded @databricks and @anyscalecompute - and now chairs Arena, the internet's default scoreboard for AI models. 0:00 The AI memory squeeze 2:55 Brain2Qwerty: Meta Reads Your Mind 9:38 Anthropic Releases Claude Science 15:58 Why Every Lab Wants Pharma 20:52 How Chatbot Arena Was Born 26:58 How Does Arena Actually Rank AI Models? 31:11 Subjective vs Objective AI Quality 33:49 How Arena Evaluates AI Agents 37:54 How Does Arena Stay Neutral? 43:31 Can AI Become Unmeasurable? @Machiz @lightspeedvp
English
Lightspeed retweetledi
would you let AI read your mind?
not your private thoughts. not yet.
Meta just released Brain2Qwerty v2, a non-invasive brain-computer interface that can decode what someone is trying to type from brain activity.
the privacy questions are obvious.
but for people who can’t speak or type, including ALS patients and stroke survivors, the access upside is enormous.
right now, the hardware is still massive, expensive, and nowhere near consumer-ready.
but if this eventually shrinks into something closer to smart glasses, it could change how we communicate with machines.
this week on Lightwork, @clairejyz and i unpack what Meta’s release means, what it doesn’t mean yet, and why the next interface may not be a keyboard, screen, or voice assistant.
we also spoke with @istoica05 about @arena, its rise to $100M in annualized run-rate revenue in just 8 months, and how it stays neutral when the labs it ranks are also its customers.
tune in below.
Claire Zau@clairejyz
NEW EPISODE: @Meta reads your brain with Brain2Qwerty. @AnthropicAI releases Claude Science, and Co-founder and Chairman of @arena Ion Stoica joins us. @istoica05 is the Berkeley professor behind an absurd run of systems that basically run modern AI: @ApacheSpark (the standard for big data), Ray (what OpenAI used to train ChatGPT), and @vllm_project (the most-used LLM inference engine). He also co-founded @databricks and @anyscalecompute - and now chairs Arena, the internet's default scoreboard for AI models. 0:00 The AI memory squeeze 2:55 Brain2Qwerty: Meta Reads Your Mind 9:38 Anthropic Releases Claude Science 15:58 Why Every Lab Wants Pharma 20:52 How Chatbot Arena Was Born 26:58 How Does Arena Actually Rank AI Models? 31:11 Subjective vs Objective AI Quality 33:49 How Arena Evaluates AI Agents 37:54 How Does Arena Stay Neutral? 43:31 Can AI Become Unmeasurable? @Machiz @lightspeedvp
English