Anyone who has worked with @JenniferHli knows she's the GOAT.
Sarah Wang 0 0 0 17 1.8K 2
Marco Mascorro
@Mascobot
Researching LLMs | Roboticist | Prev: Partner @a16z (inv @cursor_ai, @thinkymachines, @bfl_ml, etc), cofounded Fellow Robots, @BMW Research
19.5KFollowers
2.7KFollowing
6.9KPosts
815Media
16.3KLikes
Diogo Almeida@CompleteSkeptic · 22hlaunched a model
accidentally served trillions of tokens a day
29.4% of the fortune 500 showed up
raised a really big series A from @a16z
it's been 3 weeks
we would like to sleep now
accidentally served trillions of tokens a day
29.4% of the fortune 500 showed up
raised a really big series A from @a16z
it's been 3 weeks
we would like to sleep now
2:26 0 0 0 5 1K 1
I love mcmaster carr, and their website is top notch.
Can't believe they did a great site without AI even back in time.
Can't believe they did a great site without AI even back in time.
1 0 0 6 1.1K 1
The fact that many folks/investors are surprised by data and RL environments Cos that are reaching high eight-figure to mid nine-figure run rates in less than a year, tells me how late they are to the game and how little they understand about models, training and data.
11 7 2 289 51.4K 122
@Mascobot reposted
Wrote on how we're thinking about building RL environments for superintelligence
15 31 6 446 20.2K 428
@Mascobot reposted
Today we're releasing Underdog Saluki 27B
Based on Qwen 3.8 27B, Saluki is almost 7x smaller than its full-precision counterpart while it keeps 96% of its benchmark performance and beats Qwen at tool calling
It's the best model <8GB that runs in Underdog
Saluki is available today under Apache 2.0. Try today in Underdog underdog.ai/saluki
Based on Qwen 3.8 27B, Saluki is almost 7x smaller than its full-precision counterpart while it keeps 96% of its benchmark performance and beats Qwen at tool calling
It's the best model <8GB that runs in Underdog
Saluki is available today under Apache 2.0. Try today in Underdog underdog.ai/saluki
175 360 108 4K 554K 5.3K
This is very cool
MTS@MTSlive · Oct 8SITUATION DETECTED: A researcher used Claude Code with Opus 5.5 and Fable 5.1 to discover a new planet candidate by writing more than 1,000 scripts to search through seven years of NASA telescope data. NASA has approved his request for further verification.
Open quoted post → 2 1 0 10 2.4K 2
Congrats @chem_safety & @Ning_Catsnail on the launch of 🥕Karotte!
Building robust and secure RL environments is more important than ever these days because the current bottleneck on AI progress is misalignment, and developing aligned models requires robust, correct reward functions.
They have been using it to build MLE tasks, and it's been hardened through 1M+ evaluation runs and red-teaming:
Building robust and secure RL environments is more important than ever these days because the current bottleneck on AI progress is misalignment, and developing aligned models requires robust, correct reward functions.
They have been using it to build MLE tasks, and it's been hardened through 1M+ evaluation runs and red-teaming:
Jennifer Zhou@chem_safety · Oct 7Today we are open-sourcing 🥕Karotte, our framework for building RL environments.
We've used it for the past year to build MLE RL environments for frontier labs, and it's been hardened through 1M+ evaluation runs and red-teaming.
We've used it for the past year to build MLE RL environments for frontier labs, and it's been hardened through 1M+ evaluation runs and red-teaming.
Open quoted post →
3 3 0 23 2.5K 3
@Mascobot reposted
Sigil Wen, backed by a Silicon Valley's who-who, has built an on-device AI assistant that promises to be free, fully private, and capable for everyday tasks.
spr.ly/6011BGFgL7
spr.ly/6011BGFgL7
32 25 8 242 92.7K 108
@Mascobot reposted
basically any company that works goes from being a product challenge to a capital allocation challenge at scale
it is borderline impossible to find a generational CEO that isn’t a generational allocator, it’s the main function of the job. vision is overrated
it is borderline impossible to find a generational CEO that isn’t a generational allocator, it’s the main function of the job. vision is overrated
Parmita Mishra 🔬@parmita · Oct 5I will die on this hill: you simply cannot be a great founder and great investor at the same time.
The way you think about everything in life as a great investor is orthogonal to being a great founder.
You don’t have an unbiased objective view of the world as a founder.
Open quoted post →The way you think about everything in life as a great investor is orthogonal to being a great founder.
You don’t have an unbiased objective view of the world as a founder.
31 33 5 791 44.8K 321
Chris Barber@chrisbarber · Oct 2I asked people which neolabs under $10B have the highest talent density:
32 votes: Core Automation (@MillionInt, @_arohan_)
26 votes: Periodic Labs (@LiamFedus, @ekindogus)
20 votes: Flapping Airplanes (@spectorb, @amspector100, @aidanmantine)
18 votes: Standard Intelligence (@G413N, @devanshpandey)
8 votes: Ricursive (@annadgoldie, @Azaliamirh)
5 votes each:
- Ineffable (David Silver)
- Mirendil (@bneyshabur, @HarshMeh1a, @shayan_, @tararezaeikh)
- Recursive (@RichardSocher, @_rockt, @timshi_ai, @josh_tobin_, @CaimingXiong, @jeffclune, @tydsh, Alexey Dosovitskiy)
3 votes each:
- Applied Compute (@ypatil125, @rhythmrg, @lindensli)
- Isara (@ezhang7423, @hegasz)
- Prime Intellect (@vincentweisser, @johannes_hage)
2 votes each:
- Elorian (@AndrewDai, @yinfeiy, @SethInternet)
- Goodfire (@eric_ho, @DanJBalsam, @banburismus_)
- Inherent (@tantumscollins, @edwardfhughes, @LouisKirschAI, @kallyaleksiev)
- Trajectory (@rronak_, @QuantumArjun, @MichaelElabd)
- World Labs (@drfeifei, @jcjohnss, @BenMildenhall, @chlassner)
People couldn't pick a company they work at or founded. Disclosure: I'm a small investor in Applied Compute, Factory, Standard Intelligence, Trajectory and Wafer. I didn't vote.
Continued:
32 votes: Core Automation (@MillionInt, @_arohan_)
26 votes: Periodic Labs (@LiamFedus, @ekindogus)
20 votes: Flapping Airplanes (@spectorb, @amspector100, @aidanmantine)
18 votes: Standard Intelligence (@G413N, @devanshpandey)
8 votes: Ricursive (@annadgoldie, @Azaliamirh)
5 votes each:
- Ineffable (David Silver)
- Mirendil (@bneyshabur, @HarshMeh1a, @shayan_, @tararezaeikh)
- Recursive (@RichardSocher, @_rockt, @timshi_ai, @josh_tobin_, @CaimingXiong, @jeffclune, @tydsh, Alexey Dosovitskiy)
3 votes each:
- Applied Compute (@ypatil125, @rhythmrg, @lindensli)
- Isara (@ezhang7423, @hegasz)
- Prime Intellect (@vincentweisser, @johannes_hage)
2 votes each:
- Elorian (@AndrewDai, @yinfeiy, @SethInternet)
- Goodfire (@eric_ho, @DanJBalsam, @banburismus_)
- Inherent (@tantumscollins, @edwardfhughes, @LouisKirschAI, @kallyaleksiev)
- Trajectory (@rronak_, @QuantumArjun, @MichaelElabd)
- World Labs (@drfeifei, @jcjohnss, @BenMildenhall, @chlassner)
People couldn't pick a company they work at or founded. Disclosure: I'm a small investor in Applied Compute, Factory, Standard Intelligence, Trajectory and Wafer. I didn't vote.
Continued:
Open quoted post →
1 3 0 15 2.2K 5
This is a great convo and a must listen with @_sholtodouglas & @the_marwell:
Joe Lonsdale@JTLonsdale · Oct 2Some of my friends will be mad I recorded this, and comms people at Anthropic objected to releasing certain parts (and delayed it).
But it's important for leaders to make these conversations happen. And I have a lot of respect for these two.
Over a month ago, I sat down with Anthropic's key technical leaders @_sholtodouglas & @the_marwell for an optimistic insiders’ view of the AI frontier, with hard questions too: open source, regulatory capture, slowing down USA vs China, and more.
Open quoted post →But it's important for leaders to make these conversations happen. And I have a lot of respect for these two.
Over a month ago, I sat down with Anthropic's key technical leaders @_sholtodouglas & @the_marwell for an optimistic insiders’ view of the AI frontier, with hard questions too: open source, regulatory capture, slowing down USA vs China, and more.
1 0 0 24 2K 8
I’ve been using @underdogai and it’s pretty neat. It runs all locally and it’s the best private/personal AI I’ve tested with my email.
Kudos to @0xSigil & team for they launch:
Kudos to @0xSigil & team for they launch:
Sigil Wen@0xSigil · Oct 2Introducing Underdog, your Private Personal AI on devices you already own
Today we're announcing our backing from @a16z @khoslaventures @HummingbirdVC Anthology (@AnthropicAI @MenloVentures) @patrickc @naval @rauchg @polynoamial @Thom_Wolf @Mascobot @tszzl @OfficialLoganK and other top AI leaders
Our mission is to provide free, capable, reliable and private AI to billions of people.
Underdog’s Law: today’s frontier intelligence reaches your devices in six months.
We’re starting with fast, capable AI that runs on consumer hardware. By co-designing models and inference engines, we’re pushing the frontier of capability, speed, power and data efficiency.
Our research also spans agentic commerce, confidential inference, and how AI will reshape the internet economy.
Privacy and capability no longer needs to be a tradeoff. If you believe in this future, join us. Time to build.
Today we're announcing our backing from @a16z @khoslaventures @HummingbirdVC Anthology (@AnthropicAI @MenloVentures) @patrickc @naval @rauchg @polynoamial @Thom_Wolf @Mascobot @tszzl @OfficialLoganK and other top AI leaders
Our mission is to provide free, capable, reliable and private AI to billions of people.
Underdog’s Law: today’s frontier intelligence reaches your devices in six months.
We’re starting with fast, capable AI that runs on consumer hardware. By co-designing models and inference engines, we’re pushing the frontier of capability, speed, power and data efficiency.
Our research also spans agentic commerce, confidential inference, and how AI will reshape the internet economy.
Privacy and capability no longer needs to be a tradeoff. If you believe in this future, join us. Time to build.
Open quoted post →
8 6 1 47 14.2K 16
I'll be at IROS this week - ping me if you want to chat about robots, VLAs, actuators & simulation
0 0 1 9 1.4K 1
Congrats @alex_conneau & team! It was an honor to work with this team.
Alexis Conneau@alex_conneau · Sep 23Big update: @WaveFormsAI was acquired by @Meta.
Today at #MetaConnect, you’ll see a little of what we’ve all been building. ✨🤖
Open quoted post →Today at #MetaConnect, you’ll see a little of what we’ve all been building. ✨🤖
0 0 0 5 2.5K 0
Riley Walz@rtwlz · Sep 14Some news: 5 months ago I joined @coreauto, a small research lab focused on new ways for models to learn. I’m also hiring four interns to do something my younger self would've loved.
It started with a pretty unusual project. We set up a handful of small online businesses, and put an agent in charge. We gave each business some initial funding and a seed idea, and off they went, trying to sell real things to real customers while we watched. It’s about as open-ended and unforgiving a task as you can probably give to an agent, with countless decisions to make and effects that take time to become clear.
Each business got its own bank account, a debit card to pay suppliers, a web browser, an email inbox, a domain name, cloud infrastructure and a schedule. Every few hours, the agent woke up and decided what to do. It might update a landing page, adjust an ad campaign, reply to a customer, and then go back to sleep.
The agents built actual storefronts and apps. They bought ads, attracted visitors, and even made some sales. The agents acted as operators that kept a business moving, but we noticed they never really stepped back to strategize about what might move the needle. If people weren’t buying, why? Was the landing page unappealing? Or maybe the offer itself was the problem?
One amusing example was the very first order a customer placed. Our supplier had run out of stock. But it’s a dropshipping business, and thankfully tens of other suppliers carried the same product. Instead of trying another supplier, the agent almost immediately sent the customer a polite apology and attempted to issue a refund. Gah! So close!
The agents could handle much of the work of getting a business up and running. What they struggled with was deciding what to change when it wasn’t working.
That’s what we want to understand. We also want to see what changes when people take the lead. What does a business operator actually do all day when agents can do so much? What new work does this allow them to do? Which decisions can’t be delegated?
—
So this winter we're hiring 4 interns as Associate General Managers. Associate Product Manager internships are programs that give early career generalists a product to own. We want to try something similar, but with entire businesses.
You’ll create and grow businesses, with agents handling the routine work so you can focus on figuring out what matters. Each AGM gets a budget and our agent infrastructure that’s designed for this. Which ideas to try is up to you. Put more money behind what finds customers and keep refining.
Where you step in tells us what agents can't do yet. You'll be in our SF office with me and people I’m so fortunate to work with, like @_arohan_, @joannejang, @MillionInt, @marksaroufim and @juliacvillagra
We don't care much where you went to school or what your resume looks like. We’re looking for enterprising people with a drive to create, who’ve made money on the internet before and already use AI extensively to enable their ideas. If that sounds like you: coreauto.com/AGM
Open quoted post →It started with a pretty unusual project. We set up a handful of small online businesses, and put an agent in charge. We gave each business some initial funding and a seed idea, and off they went, trying to sell real things to real customers while we watched. It’s about as open-ended and unforgiving a task as you can probably give to an agent, with countless decisions to make and effects that take time to become clear.
Each business got its own bank account, a debit card to pay suppliers, a web browser, an email inbox, a domain name, cloud infrastructure and a schedule. Every few hours, the agent woke up and decided what to do. It might update a landing page, adjust an ad campaign, reply to a customer, and then go back to sleep.
The agents built actual storefronts and apps. They bought ads, attracted visitors, and even made some sales. The agents acted as operators that kept a business moving, but we noticed they never really stepped back to strategize about what might move the needle. If people weren’t buying, why? Was the landing page unappealing? Or maybe the offer itself was the problem?
One amusing example was the very first order a customer placed. Our supplier had run out of stock. But it’s a dropshipping business, and thankfully tens of other suppliers carried the same product. Instead of trying another supplier, the agent almost immediately sent the customer a polite apology and attempted to issue a refund. Gah! So close!
The agents could handle much of the work of getting a business up and running. What they struggled with was deciding what to change when it wasn’t working.
That’s what we want to understand. We also want to see what changes when people take the lead. What does a business operator actually do all day when agents can do so much? What new work does this allow them to do? Which decisions can’t be delegated?
—
So this winter we're hiring 4 interns as Associate General Managers. Associate Product Manager internships are programs that give early career generalists a product to own. We want to try something similar, but with entire businesses.
You’ll create and grow businesses, with agents handling the routine work so you can focus on figuring out what matters. Each AGM gets a budget and our agent infrastructure that’s designed for this. Which ideas to try is up to you. Put more money behind what finds customers and keep refining.
Where you step in tells us what agents can't do yet. You'll be in our SF office with me and people I’m so fortunate to work with, like @_arohan_, @joannejang, @MillionInt, @marksaroufim and @juliacvillagra
We don't care much where you went to school or what your resume looks like. We’re looking for enterprising people with a drive to create, who’ve made money on the internet before and already use AI extensively to enable their ideas. If that sounds like you: coreauto.com/AGM
0 0 0 20 2.4K 2
The reality is that all the current harnesses and abstractions are eventually going to be “eaten” by SOTA models (and their distilled counterparts if you care about speed, like Haiku, etc) just like custom prompting was back in the day, or even custom-trained fine tuned models.
@levelsio@levelsio · Sep 13I just saw this @ycombinator's batch literally only has harness startups or hardware startups
So the proof is in the pudding
In the most early adopting tech pioneering place that's Silicon Valley new startups aren't even building software anymore
And it's debatable if anyone will actually need a custom harness or it won't just be generically offered by the AI frontier companies
So software is mostly dead and hardware it is
Open quoted post →So the proof is in the pudding
In the most early adopting tech pioneering place that's Silicon Valley new startups aren't even building software anymore
And it's debatable if anyone will actually need a custom harness or it won't just be generically offered by the AI frontier companies
So software is mostly dead and hardware it is
2 0 3 12 2.6K 3
I have been testing Underdog for the last few days with my email, and this is pretty neat. This is like the first AI email app that has a lot built in out of the box, and it's all local:
Sigil Wen@0xSigil · Sep 12Anyone can get my friend’s real SSN from Claude
I'm Terrified of my data in training sets. Emails in databases. iMessages harvested. Cant trust the cloud as AI is crazy good at hacking
So I built Underdog for myself: on-device AI OS. Capable & 100% local. now my friends love it
I'm Terrified of my data in training sets. Emails in databases. iMessages harvested. Cant trust the cloud as AI is crazy good at hacking
So I built Underdog for myself: on-device AI OS. Capable & 100% local. now my friends love it
0:30 1 4 0 23 5.3K 2
The most important thing I learned from Ray Kurzweil in the early days of Singularity U is how bad the human brain is at understanding exponential growth and rates of change, and how bad humans are at extrapolating from them. He was right.
The number one thing people have consistently gotten wrong about AI model capabilities is that they look at a snapshot of current capabilities rather than the rate of change and progress.
The number one thing people have consistently gotten wrong about AI model capabilities is that they look at a snapshot of current capabilities rather than the rate of change and progress.
Felix Rieseberg@felixrieseberg · Sep 11I don’t think people have really internalized that models will continue to get dramatically better.
Two years ago, the models barely did grade school Math. Today, we’re debating which unsolvable Math problems will be tackled. That’s nuts! Of course you sound silly saying that more capability improvements are coming our way at an even faster rate than before. Even if people believe us, it’s hard for many MTS to really wrap their heads around the implications.
In related news, I recommend reading this threat report we published today. It’s comprehensive and hopefully food for more thought - combine the expected rate and magnitude of improvement with threats for a better insight into how many engineers and researchers at the frontier feel.
anthropic.com/threat-intelli…
Open quoted post →Two years ago, the models barely did grade school Math. Today, we’re debating which unsolvable Math problems will be tackled. That’s nuts! Of course you sound silly saying that more capability improvements are coming our way at an even faster rate than before. Even if people believe us, it’s hard for many MTS to really wrap their heads around the implications.
In related news, I recommend reading this threat report we published today. It’s comprehensive and hopefully food for more thought - combine the expected rate and magnitude of improvement with threats for a better insight into how many engineers and researchers at the frontier feel.
anthropic.com/threat-intelli…
43 63 6 607 43.6K 127
Codex computer use is just very good now and fast. I'm impressed on how much it improved in the last ~6 months.
Great job @thsottiaux, @AriX & Team.
Great job @thsottiaux, @AriX & Team.
2 0 0 9 1.5K 0