Electric Plant CEO @justinmast reveals how his team built a Neuralink for trees that can detect when plants are thirsty or stressed by reading their electrical signals:
"Us and a very few other people in the world are recording these signals, and we’re training models to see physiology of the plant."
"With this first model release, we can see the photosystem efficiency, how well is the plant turning sunlight into sugar, in real time. We can see if the plant’s thirsty, if it’s stressed."
"Right now it’s functioning like a WHOOP band for a tree. But we’re tapped in directly to those electrical signals, so it functions like a Neuralink for trees, too."
@electricplantco
"Us and a very few other people in the world are recording these signals, and we’re training models to see physiology of the plant."
"With this first model release, we can see the photosystem efficiency, how well is the plant turning sunlight into sugar, in real time. We can see if the plant’s thirsty, if it’s stressed."
"Right now it’s functioning like a WHOOP band for a tree. But we’re tapped in directly to those electrical signals, so it functions like a Neuralink for trees, too."
@electricplantco
1/ Today we’re unveiling Bombadil, @electricplantco's foundation model that reads the electrical signals of living plants and turns them into a continuous view of their physiology.
Trained on 7M+ hours of plant electrical signal across 125+ species. That makes Bombadil the most f directly recorded electrical biosignals in existence. The animal equivalent would be one model that reads the brainwaves of blue whales, horseshoe crabs and fruit flies.
4
4
73
Electric Plant CEO @justinmast explains how plants can hear insects approaching and change their chemistry to defend themselves before being attacked:
"What we can do now is decode eight different vital signs or physiological functions inside the plant. That’s what we’re going what’s in the paper."
"On top of that, there’s this whole list of plant capabilities. Peer-reviewed science has shown plants are capable of hearing insects come."
"They’re capable of changing their chemistry to deter those insects in advance. They’re communicating with each other. They’re acutely aware of the chemical composition in the air or in the soil."
"We keep a list that’s probably up to 78 things. That’s our product roadmap. Some of that will work, and some of those things won’t work, too."
@electricplantco
"What we can do now is decode eight different vital signs or physiological functions inside the plant. That’s what we’re going what’s in the paper."
"On top of that, there’s this whole list of plant capabilities. Peer-reviewed science has shown plants are capable of hearing insects come."
"They’re capable of changing their chemistry to deter those insects in advance. They’re communicating with each other. They’re acutely aware of the chemical composition in the air or in the soil."
"We keep a list that’s probably up to 78 things. That’s our product roadmap. Some of that will work, and some of those things won’t work, too."
@electricplantco
1/ Today we’re unveiling Bombadil, @electricplantco's foundation model that reads the electrical signals of living plants and turns them into a continuous view of their physiology.
Trained on 7M+ hours of plant electrical signal across 125+ species. That makes Bombadil the most f directly recorded electrical biosignals in existence. The animal equivalent would be one model that reads the brainwaves of blue whales, horseshoe crabs and fruit flies.
4
2
47
Electric Plant CEO @justinmast reveals how plants' electrical signals can tell exactly when someone enters a room:
"Our AI team was like, hey Justin, we can tell when you’re coming into the room and when you’re leaving based purely on the electrical signals of plants."
"I was at’s true. They’re like, yesterday you walked in at 7:53 AM, the day before that, 8:03 AM."
"It’s this weird combination of their potential sensing of us, door opening and closing changes the pressure signature in the room, lights on and off, plants are acutely aware of the light levels. All those things combined, they had pinned me down."
@electricplantco
"Our AI team was like, hey Justin, we can tell when you’re coming into the room and when you’re leaving based purely on the electrical signals of plants."
"I was at’s true. They’re like, yesterday you walked in at 7:53 AM, the day before that, 8:03 AM."
"It’s this weird combination of their potential sensing of us, door opening and closing changes the pressure signature in the room, lights on and off, plants are acutely aware of the light levels. All those things combined, they had pinned me down."
@electricplantco
1/ Today we’re unveiling Bombadil, @electricplantco's foundation model that reads the electrical signals of living plants and turns them into a continuous view of their physiology.
Trained on 7M+ hours of plant electrical signal across 125+ species. That makes Bombadil the most f directly recorded electrical biosignals in existence. The animal equivalent would be one model that reads the brainwaves of blue whales, horseshoe crabs and fruit flies.
2
0
25
American Housing CEO @rileymeik explains how AI could make it possible to mass-produce thousands of homes a year without making them identical:
"Most manufactured products built in high volume, you’re building the same thing over and over. All iPhones are the same, all Model Ys are different. We have to adapt to different sites, different zoning requirements, different building codes."
"Without AI, we would have to throw bodies at this, tons of architects and engineers designing custom homes for every site."
"Today, we have a standardized kit of parts, think of them as Lego pieces. The same walls, they just come in different sizes. We use AI to figure out the ideal way to arrange these components into the best house on every site."
"At some point the company will be building thousands and thousands of homes a year with very few of those being identical to each other, which is an incredibly hard problem when you’re trying to manufacture at scale. AI really just makes that possible."
@americanhousing
"Most manufactured products built in high volume, you’re building the same thing over and over. All iPhones are the same, all Model Ys are different. We have to adapt to different sites, different zoning requirements, different building codes."
"Without AI, we would have to throw bodies at this, tons of architects and engineers designing custom homes for every site."
"Today, we have a standardized kit of parts, think of them as Lego pieces. The same walls, they just come in different sizes. We use AI to figure out the ideal way to arrange these components into the best house on every site."
"At some point the company will be building thousands and thousands of homes a year with very few of those being identical to each other, which is an incredibly hard problem when you’re trying to manufacture at scale. AI really just makes that possible."
@americanhousing
For a half-century, America has critically under-built family-sized housing in our most dynamic cities and neighborhoods, rendering them childless and unaffordable.
It's time to save the American Dream.
Introducing The American Housing Corporation.
4
1
36
American Housing CEO @rileymeik explains why prefab housing never revolutionized construction, and how he's redesigning homes to be built like cars:
"Prefab has been around for a very long time, but typically housing manufacturers produce parts that they sell to home builders. ing two-by-fours and drywall."
"Because we’re vertically integrated, we get to choose what we build and how we build it. We’re not using the hand tools you would buy from Home Depot. We’re using the same equipment you would see in an automotive or aerospace factory."
"This is the first house built using a system that’s meant to be produced by machines and put together by machines rather than built by hand by people."
@americanhousing
"Prefab has been around for a very long time, but typically housing manufacturers produce parts that they sell to home builders. ing two-by-fours and drywall."
"Because we’re vertically integrated, we get to choose what we build and how we build it. We’re not using the hand tools you would buy from Home Depot. We’re using the same equipment you would see in an automotive or aerospace factory."
"This is the first house built using a system that’s meant to be produced by machines and put together by machines rather than built by hand by people."
@americanhousing
For a half-century, America has critically under-built family-sized housing in our most dynamic cities and neighborhoods, rendering them childless and unaffordable.
It's time to save the American Dream.
Introducing The American Housing Corporation.
0
2
32
Matter Labs CEO @gluk64 warns AI and quantum computing are exposing a dangerous flaw in how banks secure their systems:
"Commander does not require you to run a Prividium ledger. You can do it on top of your existing Besu chain, you can do it with public blockchains. From there idium Ledger, our secure high performance ZK proven chain."
"Some players in the space are very actively looking into upgrading the core banking systems, because they feel the pressure from AI and from the quantum advances, to be ready for Q day."
"A friend of mine said many years ago that Bitcoin is the best thing that happened to cybersecurity. It’s finally entering the phase where they feel in full how dangerous it is to operate systems based on security by obscurity and not based on public standards open to public scrutiny."
"Zero knowledge proofs, blockchains can help defend yourself in this new era of AI and quantum."
@zksync
"Commander does not require you to run a Prividium ledger. You can do it on top of your existing Besu chain, you can do it with public blockchains. From there idium Ledger, our secure high performance ZK proven chain."
"Some players in the space are very actively looking into upgrading the core banking systems, because they feel the pressure from AI and from the quantum advances, to be ready for Q day."
"A friend of mine said many years ago that Bitcoin is the best thing that happened to cybersecurity. It’s finally entering the phase where they feel in full how dangerous it is to operate systems based on security by obscurity and not based on public standards open to public scrutiny."
"Zero knowledge proofs, blockchains can help defend yourself in this new era of AI and quantum."
@zksync
Introducing Prividium Commander: your digital assets control plane.
Institutions can now take digital assets from pilots to revenue with one product that runs under their own rules and connects to the systems they already use.
1
3
24
SITUATION EXPLAINED: China's Kimi K3.1 Pro got so good at hacking that Moonshot stripped its cyber skills before open-sourcing it.
• Kimi K3.1 Pro has 2.8T parameters (104B active) and a 1M-token context window, with full weights due by Nov 4
• It nearly matches Opus 5.5 on 7% vs 66.4%) and beats it on SWE Marathon, HLE with tools, and OfficeQA Pro v2
• API pricing is $2/$10 per million tokens, a third cheaper than Kimi K3
• Removing cyber data from training dropped its score on real exploit tasks from 90.5% to 81%, now below DeepSeek V4.1 Flash at 85.7%
• Moonshot pledged not to release open weights that "extend the absolute frontier of cyber capability"
• It still plans to raise the cyber capability of its open weights over time, saying defenders need dual-use tools too
@theojaffee: "I don’t think I’ve ever heard of a Chinese model or a Chinese lab nerfing any of their models on anything ever."
• Kimi K3.1 Pro has 2.8T parameters (104B active) and a 1M-token context window, with full weights due by Nov 4
• It nearly matches Opus 5.5 on 7% vs 66.4%) and beats it on SWE Marathon, HLE with tools, and OfficeQA Pro v2
• API pricing is $2/$10 per million tokens, a third cheaper than Kimi K3
• Removing cyber data from training dropped its score on real exploit tasks from 90.5% to 81%, now below DeepSeek V4.1 Flash at 85.7%
• Moonshot pledged not to release open weights that "extend the absolute frontier of cyber capability"
• It still plans to raise the cyber capability of its open weights over time, saying defenders need dual-use tools too
@theojaffee: "I don’t think I’ve ever heard of a Chinese model or a Chinese lab nerfing any of their models on anything ever."
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1…
11
15
222
General Reasoning CEO @rosstaylor90 reveals how their new RLCR experiment led AI agents running a simulated company to independently hire middle managers:
"My background was in economics. How do you actually test these things? You don’t have natural experiments. The beauty of ives is you can actually test these things."
"We had a CEO agent and some specialists, and just said, you decide who to hire. They’d hire lead engineers, middle management. I’m like, oh my God, is this really AGI, middle management? But that is actually predicted by economic theory."
"The bigger question is where does the human fit in. In a lot of these domains you can probably go fully automated, but is that even desirable? If I had an automated hedge fund, I’m sure I’d want some say at some level of decision-making."
@GenReasoning
"My background was in economics. How do you actually test these things? You don’t have natural experiments. The beauty of ives is you can actually test these things."
"We had a CEO agent and some specialists, and just said, you decide who to hire. They’d hire lead engineers, middle management. I’m like, oh my God, is this really AGI, middle management? But that is actually predicted by economic theory."
"The bigger question is where does the human fit in. In a lot of these domains you can probably go fully automated, but is that even desirable? If I had an automated hedge fund, I’m sure I’d want some say at some level of decision-making."
@GenReasoning
Introducing Reinforcement Learning from Collective Rewards (RLCR).
RLCR is a recipe that trains language models as agent organisations, which share a message board and workspace. We use it to train Avalon-M1-Mini (120B) and study its behaviour.
We find a number of interesting Learnt organisational hierarchies and divisions of labour
🕵️ Increased collective verification, through hires of verifiers and testers
👥 The rise of collective thinking, where agents think more about other agents and how to help them
We believe RLCR can provide a laboratory for studying collective intelligence and multi-agent safety.
Blog post and technical report 👇
4
2
40
General Reasoning CEO @rosstaylor90 explains how their new RLCR experiment taught AI agents to organize themselves, hire teammates, and decide when to DM each other:
"The idea of RLCR is, instead of just trying to optimize single agents, how do you optimize collective arch communities, and eventually economies, markets."
"We optimize an organization in a Slack-like harness. They learn who they should hire, what roles they should hire for, when they should be pinging someone annoyingly in their direct messages versus making a public channel."
"It’s not just another model, it’s a whole new ball game. It’s a whole new direction."
@GenReasoning
"The idea of RLCR is, instead of just trying to optimize single agents, how do you optimize collective arch communities, and eventually economies, markets."
"We optimize an organization in a Slack-like harness. They learn who they should hire, what roles they should hire for, when they should be pinging someone annoyingly in their direct messages versus making a public channel."
"It’s not just another model, it’s a whole new ball game. It’s a whole new direction."
@GenReasoning
Introducing Reinforcement Learning from Collective Rewards (RLCR).
RLCR is a recipe that trains language models as agent organisations, which share a message board and workspace. We use it to train Avalon-M1-Mini (120B) and study its behaviour.
We find a number of interesting Learnt organisational hierarchies and divisions of labour
🕵️ Increased collective verification, through hires of verifiers and testers
👥 The rise of collective thinking, where agents think more about other agents and how to help them
We believe RLCR can provide a laboratory for studying collective intelligence and multi-agent safety.
Blog post and technical report 👇
3
3
22
Marathon Fusion CEO @kylecschiller explains why superintelligence could do far more for fusion than for solar panels:
"There’s all this discourse around what we’re going to do in the future when we have superintelligence. We’re probably not going to optimize solar panels that ready very close to the maximally efficient end state of what a solar panel is."
"I think about how bad fusion is today, and I get really excited instead of depressed."
"We have this machine that the more intelligence we throw at it, the better it gets at spitting out energy, and also all other forms of matter. That’s the thing I wanna bet on."
@jfp_fusioneer @MarathonFusion
"There’s all this discourse around what we’re going to do in the future when we have superintelligence. We’re probably not going to optimize solar panels that ready very close to the maximally efficient end state of what a solar panel is."
"I think about how bad fusion is today, and I get really excited instead of depressed."
"We have this machine that the more intelligence we throw at it, the better it gets at spitting out energy, and also all other forms of matter. That’s the thing I wanna bet on."
@jfp_fusioneer @MarathonFusion
Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10.
Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years.
Now, we’re building Francisco to achieve record power density.
4
4
52
Marathon Fusion CSO @jfp_fusioneer reveals how his team turned mercury into real gold and beat a world record previously held by CERN:
"It started about 18 months ago when Adam Rutkowski, the CTO, discovered that you could go from mercury to gold with a neutron reaction. We’re f nuclides. We’re arbitraging price differences in matter."
"We wanted to beat the CERN record. CERN has this multi-billion dollar facility operated by some of the smartest people in the world."
"Maybe a scrappy fusion company can go out and beat that. We procured some natural mercury and used a very modest neutron source."
"There’s an isotope of mercury. You hit mercury with a neutron, and you knock out two neutrons. The slightly lighter mercury decays in a couple of days into stable gold. We measured the photons, calculated the rate at which we were making gold, compared it to the previous world record, and it turned out we surpassed that."
@kylecschiller @MarathonFusion
"It started about 18 months ago when Adam Rutkowski, the CTO, discovered that you could go from mercury to gold with a neutron reaction. We’re f nuclides. We’re arbitraging price differences in matter."
"We wanted to beat the CERN record. CERN has this multi-billion dollar facility operated by some of the smartest people in the world."
"Maybe a scrappy fusion company can go out and beat that. We procured some natural mercury and used a very modest neutron source."
"There’s an isotope of mercury. You hit mercury with a neutron, and you knock out two neutrons. The slightly lighter mercury decays in a couple of days into stable gold. We measured the photons, calculated the rate at which we were making gold, compared it to the previous world record, and it turned out we surpassed that."
@kylecschiller @MarathonFusion
Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10.
Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years.
Now, we’re building Francisco to achieve record power density.
2
0
19
Marathon Fusion CEO @kylecschiller explains how turning mercury into gold could make alchemy a trillion-dollar industry:
"Mercury and gold are literally one proton away. There’s one bit in there, if you can flip this, you’ve created a different element."
"Particle accelerators, ologies have existed. There’s just nothing out there that really takes us up to the trillion dollar level. Fusion will."
"People compare this to synthetic diamonds or aluminum. Aluminum prices cratered over a 100,000x increase in supply. What we’re proposing is comparatively modest."
“If things go really well, we might double global gold production. That’s another $500 billion per year that we can invest in clean energy. This is an incredibly resilient asset."
@jfp_fusioneer @MarathonFusion
"Mercury and gold are literally one proton away. There’s one bit in there, if you can flip this, you’ve created a different element."
"Particle accelerators, ologies have existed. There’s just nothing out there that really takes us up to the trillion dollar level. Fusion will."
"People compare this to synthetic diamonds or aluminum. Aluminum prices cratered over a 100,000x increase in supply. What we’re proposing is comparatively modest."
“If things go really well, we might double global gold production. That’s another $500 billion per year that we can invest in clean energy. This is an incredibly resilient asset."
@jfp_fusioneer @MarathonFusion
Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10.
Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years.
Now, we’re building Francisco to achieve record power density.
3
5
90
a16z's @YODELAGARZA predicts governments will have to shut down critical infrastructure within the next year to patch it before AI-powered hackers get in:
"I think it’s almost certain, but not for the reasons that bad stuff is gonna happen. Governments will take systems down. will just be shut off until they can patch it and bring it back safely."
"The banks that got access to Glasswing six, eight months ago, they’ve been doing nothing but patching for the last six months."
"I think it was either Bank of America or JP Morgan that’s hiring an additional thousand people to focus just on patching."
@a16z
"I think it’s almost certain, but not for the reasons that bad stuff is gonna happen. Governments will take systems down. will just be shut off until they can patch it and bring it back safely."
"The banks that got access to Glasswing six, eight months ago, they’ve been doing nothing but patching for the last six months."
"I think it was either Bank of America or JP Morgan that’s hiring an additional thousand people to focus just on patching."
@a16z
Two critical vulnerabilities have been identified in PHP, a widely used programming language that powers websites and web applications. These vulnerabilities (CVE-2026-109618 and CVE-2026-109619) could allow malicious cyber actors to remotely execute code in PHP core, potentially data and networks.
The FBI urges organizations to take the following steps now:
➡️ If an immediate update is not feasible it is recommended that organizations take the system offline until it can be fully mitigated.
➡️ Immediately update affected PHP installations following the guidance on php.net. Inventory where PHP runs in your environment, prioritizing internet-facing systems and vendor products that bundle PHP.
➡️ Monitor web servers for suspicious activity, including PHP processes unexpectedly launching command shells or downloading files. Ensure appropriate process-monitoring capabilities are in place to help identify potential exploitation.
➡️ If you find signs of compromise, isolate affected systems, preserve your web server and PHP logs, and report it to your local FBI field office or ic3.gov. Those reports help us identify the actors involved and warn other organizations before they become victims.
For the latest information, follow PHP’s announcement on php.net, CERT-EU Security Advisory 2026-018, and CISA’s Known Exploited Vulnerabilities catalog.
1
4
30
a16z's @YODELAGARZA warns AI is making the security flaws companies chose to leave unfixed easy to exploit:
"The first CISO happened at Citigroup, right after a bunch of Russian hackers got in and stole a ton of money. In the late ’90s, systems were trivially secure. You could t, connect to a port, and start issuing commands as root."
"Into the security world crept this concept of risk acceptance. The thinking was, to exploit these vulnerabilities, it’s going to take hundreds of hours of very technical people. The same way we rate physical safes, by how many hours it takes to break in."
"What’s happened in the last two, three years with this AI revolution is that the cost of that knowledge work has gone essentially to zero. We’re finding issues in code that have been there for many years just because agents and bots are finally looking."
"For the next three to six months, we’re just gonna be seeing a ton of these. We have the PHP one, one in Bind, one in the Linux kernel. There’s four or five others that are currently embargoed."
@a16z
"The first CISO happened at Citigroup, right after a bunch of Russian hackers got in and stole a ton of money. In the late ’90s, systems were trivially secure. You could t, connect to a port, and start issuing commands as root."
"Into the security world crept this concept of risk acceptance. The thinking was, to exploit these vulnerabilities, it’s going to take hundreds of hours of very technical people. The same way we rate physical safes, by how many hours it takes to break in."
"What’s happened in the last two, three years with this AI revolution is that the cost of that knowledge work has gone essentially to zero. We’re finding issues in code that have been there for many years just because agents and bots are finally looking."
"For the next three to six months, we’re just gonna be seeing a ton of these. We have the PHP one, one in Bind, one in the Linux kernel. There’s four or five others that are currently embargoed."
@a16z
Two critical vulnerabilities have been identified in PHP, a widely used programming language that powers websites and web applications. These vulnerabilities (CVE-2026-109618 and CVE-2026-109619) could allow malicious cyber actors to remotely execute code in PHP core, potentially data and networks.
The FBI urges organizations to take the following steps now:
➡️ If an immediate update is not feasible it is recommended that organizations take the system offline until it can be fully mitigated.
➡️ Immediately update affected PHP installations following the guidance on php.net. Inventory where PHP runs in your environment, prioritizing internet-facing systems and vendor products that bundle PHP.
➡️ Monitor web servers for suspicious activity, including PHP processes unexpectedly launching command shells or downloading files. Ensure appropriate process-monitoring capabilities are in place to help identify potential exploitation.
➡️ If you find signs of compromise, isolate affected systems, preserve your web server and PHP logs, and report it to your local FBI field office or ic3.gov. Those reports help us identify the actors involved and warn other organizations before they become victims.
For the latest information, follow PHP’s announcement on php.net, CERT-EU Security Advisory 2026-018, and CISA’s Known Exploited Vulnerabilities catalog.
3
0
16
Dark Æon author @JOEBOTxyz challenges Ray Kurzweil's singularity timeline with the predictions that didn’t come true:
"If you read carefully the specifics of The Age of Spiritual Machines and The Singularity Is Near, it wasn’t some amazing one-to-one prediction of the future."
. The lines kept going up. But he predicted that by now most processing would be done on carbon nanotubes, most people would live within virtual reality, most homes would have a humanoid robot. None of that happened."
"Futurism is not an exact science, and there are as many predicted futures as there are fevered egos."
"I grew up with all these predictions of apocalypses. If the singularity is the apocalypse, that’s the one that pulled the right lottery number, because all the others have not yet panned out."
"If you read carefully the specifics of The Age of Spiritual Machines and The Singularity Is Near, it wasn’t some amazing one-to-one prediction of the future."
. The lines kept going up. But he predicted that by now most processing would be done on carbon nanotubes, most people would live within virtual reality, most homes would have a humanoid robot. None of that happened."
"Futurism is not an exact science, and there are as many predicted futures as there are fevered egos."
"I grew up with all these predictions of apocalypses. If the singularity is the apocalypse, that’s the one that pulled the right lottery number, because all the others have not yet panned out."
5
7
67
Dark Æon author @JOEBOTxyz asks who Altman, Amodei and Musk are really building AI for:
"You are told that something is inevitable, and that without it you will no longer be relevant."
"Accepting artificial superintelligence and the technologies that are supposed to keep us om brain chips to nanobots to genetic upgrades so we can have big swollen heads to keep up with the large language models."
"If someone tells you they are going to transform everything about your life, yourself, your community, your civilization, then you have to ask the question, for whose benefit?"
"You are told that something is inevitable, and that without it you will no longer be relevant."
"Accepting artificial superintelligence and the technologies that are supposed to keep us om brain chips to nanobots to genetic upgrades so we can have big swollen heads to keep up with the large language models."
"If someone tells you they are going to transform everything about your life, yourself, your community, your civilization, then you have to ask the question, for whose benefit?"
5
11
63
SpaceXSI's @mattyp says "UI is dead," and the next generation of internal tools won't need a UI:
"UI is dead. I just gotta throw out the hot takes. In all seriousness, I build UIs every day. But a surprising number of things that I thought I needed a UI for, I really didn’t."
hen vibe coding internal tools. There’s this next phase of internal tooling that is just chat, just an agent."
"Software is really just a UI, a back end, and a data source. If the agent acts as the back end and has access to a spreadsheet, and chat is the UI, you can get a surprising number of things done."
@bot @SpaceXSI
"UI is dead. I just gotta throw out the hot takes. In all seriousness, I build UIs every day. But a surprising number of things that I thought I needed a UI for, I really didn’t."
hen vibe coding internal tools. There’s this next phase of internal tooling that is just chat, just an agent."
"Software is really just a UI, a back end, and a data source. If the agent acts as the back end and has access to a spreadsheet, and chat is the UI, you can get a surprising number of things done."
@bot @SpaceXSI
FULL INTERVIEW: SpaceXSI's @mattyp says agents now build in an instant what used to take him a whole weekend. He still doesn't think software engineers will ever go away, because the "model doesn't really have empathy for people."
@mattyp works on @bot at @SpaceXSI. He joined wn why chat might replace most apps, how Grok Bot keeps your passwords hidden from the bot itself, and what agents still can't do:
00:56 why have multiple Grok Bots instead of one
03:52 why chat might be the only UI you need
06:31 why chat is just the first form factor for agents
07:43 how Grok Bot handles security
09:22 the Gmail alerts from a VM in Korea
10:27 treat your agent like a new hire
11:23 the cantaloupe problem with AI memory
13:45 why software engineers won't go away
15:07 where models still fall short
17:49 how SpaceXSI uses Grok Bot internally
19:33 working for @elonmusk after Cursor
21:05 what comes after the IDE
16
12
161
SpaceXSI’s @mattyp explains why software built for humans is slowing down AI agents:
"Late 2025 everybody was pretty skeptical. The discourse on X was, coding agents are cool, but I still read all the code. Then something happened 2025 to 2026 where people were like, actually I
"All of these tools around knowledge work were designed for humans over the past 30, 40 years. It’s kinda funny to think about Grok Bot, an agent with a computer that is basically using a remote desktop to do things for humans."
"Booking a flight, interacting with spreadsheets, making decks, these are designed for humans and might be better designed for agents."
@bot @SpaceXSI
"Late 2025 everybody was pretty skeptical. The discourse on X was, coding agents are cool, but I still read all the code. Then something happened 2025 to 2026 where people were like, actually I
"All of these tools around knowledge work were designed for humans over the past 30, 40 years. It’s kinda funny to think about Grok Bot, an agent with a computer that is basically using a remote desktop to do things for humans."
"Booking a flight, interacting with spreadsheets, making decks, these are designed for humans and might be better designed for agents."
@bot @SpaceXSI
FULL INTERVIEW: SpaceXSI's @mattyp says agents now build in an instant what used to take him a whole weekend. He still doesn't think software engineers will ever go away, because the "model doesn't really have empathy for people."
@mattyp works on @bot at @SpaceXSI. He joined wn why chat might replace most apps, how Grok Bot keeps your passwords hidden from the bot itself, and what agents still can't do:
00:56 why have multiple Grok Bots instead of one
03:52 why chat might be the only UI you need
06:31 why chat is just the first form factor for agents
07:43 how Grok Bot handles security
09:22 the Gmail alerts from a VM in Korea
10:27 treat your agent like a new hire
11:23 the cantaloupe problem with AI memory
13:45 why software engineers won't go away
15:07 where models still fall short
17:49 how SpaceXSI uses Grok Bot internally
19:33 working for @elonmusk after Cursor
21:05 what comes after the IDE
3
2
28
SpaceXSI’s @mattyp reveals how Grok Bot can log into your accounts without ever seeing your password:
"There’s a reviewer agent that oversees every action the bot can take. If I was gonna send an email or a Slack message or fill in sensitive information, Grok Bot is going to agent and require your permission."
"First days of personal agents, all your credentials were stored in plain text on a disk somewhere. Today that’s much different."
"When you give Grok Bot information, it’s stored securely in a vault, so both the bot and the computer don’t have access. When you put your password or credit card info into a website, that data is encrypted from the form to the browser. The bot actually doesn’t know what it is."
@bot @SpaceXSI
"There’s a reviewer agent that oversees every action the bot can take. If I was gonna send an email or a Slack message or fill in sensitive information, Grok Bot is going to agent and require your permission."
"First days of personal agents, all your credentials were stored in plain text on a disk somewhere. Today that’s much different."
"When you give Grok Bot information, it’s stored securely in a vault, so both the bot and the computer don’t have access. When you put your password or credit card info into a website, that data is encrypted from the form to the browser. The bot actually doesn’t know what it is."
@bot @SpaceXSI
FULL INTERVIEW: SpaceXSI's @mattyp says agents now build in an instant what used to take him a whole weekend. He still doesn't think software engineers will ever go away, because the "model doesn't really have empathy for people."
@mattyp works on @bot at @SpaceXSI. He joined wn why chat might replace most apps, how Grok Bot keeps your passwords hidden from the bot itself, and what agents still can't do:
00:56 why have multiple Grok Bots instead of one
03:52 why chat might be the only UI you need
06:31 why chat is just the first form factor for agents
07:43 how Grok Bot handles security
09:22 the Gmail alerts from a VM in Korea
10:27 treat your agent like a new hire
11:23 the cantaloupe problem with AI memory
13:45 why software engineers won't go away
15:07 where models still fall short
17:49 how SpaceXSI uses Grok Bot internally
19:33 working for @elonmusk after Cursor
21:05 what comes after the IDE
3
4
81
SpaceXSI's @mattyp reveals why Grok Bot users keep creating an AI chief of staff to manage their other bots:
"What we found is that almost everyone had this usage pattern where they’re creating a chief of staff or a single point bot that would then manage these other bots."
"We el like an AI companion such that you’re creating these little coworkers. But having one bot that serves as an inbox or a catchall is something that most people do, and we’ve started to build it into the product."
"One primary bot, and then specific bots for different use cases, which if you’re gonna create team bots or share them with folks on your team, turns out to be pretty helpful."
@bot @SpaceXSI
"What we found is that almost everyone had this usage pattern where they’re creating a chief of staff or a single point bot that would then manage these other bots."
"We el like an AI companion such that you’re creating these little coworkers. But having one bot that serves as an inbox or a catchall is something that most people do, and we’ve started to build it into the product."
"One primary bot, and then specific bots for different use cases, which if you’re gonna create team bots or share them with folks on your team, turns out to be pretty helpful."
@bot @SpaceXSI
FULL INTERVIEW: SpaceXSI's @mattyp says agents now build in an instant what used to take him a whole weekend. He still doesn't think software engineers will ever go away, because the "model doesn't really have empathy for people."
@mattyp works on @bot at @SpaceXSI. He joined wn why chat might replace most apps, how Grok Bot keeps your passwords hidden from the bot itself, and what agents still can't do:
00:56 why have multiple Grok Bots instead of one
03:52 why chat might be the only UI you need
06:31 why chat is just the first form factor for agents
07:43 how Grok Bot handles security
09:22 the Gmail alerts from a VM in Korea
10:27 treat your agent like a new hire
11:23 the cantaloupe problem with AI memory
13:45 why software engineers won't go away
15:07 where models still fall short
17:49 how SpaceXSI uses Grok Bot internally
19:33 working for @elonmusk after Cursor
21:05 what comes after the IDE
6
4
34
SpaceXSI's @mattyp explains why chatting with AI agents today is like using the original iPhone:
"Chat is not the final form factor for these personal assistants. It’s great, but think back to when the iPhone first came out, there’s skeuomorphic design."
"Everything had to look eople weren’t used to digital interfaces. Now everything’s flat because people know how to interact with these things."
"Maybe chat is the first form factor of interacting with agents, and I think we’re very early on in that."
@SpaceXSI @bot
"Chat is not the final form factor for these personal assistants. It’s great, but think back to when the iPhone first came out, there’s skeuomorphic design."
"Everything had to look eople weren’t used to digital interfaces. Now everything’s flat because people know how to interact with these things."
"Maybe chat is the first form factor of interacting with agents, and I think we’re very early on in that."
@SpaceXSI @bot
FULL INTERVIEW: SpaceXSI's @mattyp says agents now build in an instant what used to take him a whole weekend. He still doesn't think software engineers will ever go away, because the "model doesn't really have empathy for people."
@mattyp works on @bot at @SpaceXSI. He joined wn why chat might replace most apps, how Grok Bot keeps your passwords hidden from the bot itself, and what agents still can't do:
00:56 why have multiple Grok Bots instead of one
03:52 why chat might be the only UI you need
06:31 why chat is just the first form factor for agents
07:43 how Grok Bot handles security
09:22 the Gmail alerts from a VM in Korea
10:27 treat your agent like a new hire
11:23 the cantaloupe problem with AI memory
13:45 why software engineers won't go away
15:07 where models still fall short
17:49 how SpaceXSI uses Grok Bot internally
19:33 working for @elonmusk after Cursor
21:05 what comes after the IDE
5
9
55
FULL INTERVIEW: SpaceXSI's @mattyp says agents now build in an instant what used to take him a whole weekend. He still doesn't think software engineers will ever go away, because the "model doesn't really have empathy for people."
@mattyp works on @bot at @SpaceXSI. He joined wn why chat might replace most apps, how Grok Bot keeps your passwords hidden from the bot itself, and what agents still can't do:
00:56 why have multiple Grok Bots instead of one
03:52 why chat might be the only UI you need
06:31 why chat is just the first form factor for agents
07:43 how Grok Bot handles security
09:22 the Gmail alerts from a VM in Korea
10:27 treat your agent like a new hire
11:23 the cantaloupe problem with AI memory
13:45 why software engineers won't go away
15:07 where models still fall short
17:49 how SpaceXSI uses Grok Bot internally
19:33 working for @elonmusk after Cursor
21:05 what comes after the IDE
@mattyp works on @bot at @SpaceXSI. He joined wn why chat might replace most apps, how Grok Bot keeps your passwords hidden from the bot itself, and what agents still can't do:
00:56 why have multiple Grok Bots instead of one
03:52 why chat might be the only UI you need
06:31 why chat is just the first form factor for agents
07:43 how Grok Bot handles security
09:22 the Gmail alerts from a VM in Korea
10:27 treat your agent like a new hire
11:23 the cantaloupe problem with AI memory
13:45 why software engineers won't go away
15:07 where models still fall short
17:49 how SpaceXSI uses Grok Bot internally
19:33 working for @elonmusk after Cursor
21:05 what comes after the IDE
7
9
115
CFR's @ChrisRMcGuire warns what happens if China gets frontier AI first:
Its intelligence agencies won't be disclosing cyber vulnerabilities, they'll be exploiting them.
"If China develops the frontier models, that is catastrophic for national security. Then it’s not the NSA exploit. It’s not our companies developing Mythos. It’s the MSS, and they’re not gonna be alerting the world."
"The controls give us the lead to spend on safety at the right time. We didn’t think 2022 was the right time. I think it probably is now."
"It’s really not credible to claim we can’t do anything because of China while we are actively supporting China’s AI development. We’re actively fueling their tank."
Its intelligence agencies won't be disclosing cyber vulnerabilities, they'll be exploiting them.
"If China develops the frontier models, that is catastrophic for national security. Then it’s not the NSA exploit. It’s not our companies developing Mythos. It’s the MSS, and they’re not gonna be alerting the world."
"The controls give us the lead to spend on safety at the right time. We didn’t think 2022 was the right time. I think it probably is now."
"It’s really not credible to claim we can’t do anything because of China while we are actively supporting China’s AI development. We’re actively fueling their tank."
Three quick reactions to how the new Chinese export controls on AI, which give the Chinese government the authority to regulate the export, licensing, or publishing of any AI model (including open-source models), will impact the US government and industry:
(1) This introduces significant uncertainty for US businesses that host or rely on Chinese models. Particularly given that Chinese AI labs are moving toward licensing the use of open-source models to generate revenue, the Chinese government stating that it has the ability to curtail those licensing agreements at any time should make US businesses more hesitant to rely on those models going forward.
(2) This highlights why the US government also needs to adopt a framework that comprehensively governs the use of US AI technology overseas. The Chinese government clearly sees it as a national security imperative to understand who is utilizing or creating Chinese AI tech, and that it has the ability to regulate it--to include not only AI models, but also chips and data centers (even if China isn't exporting many of the latter). Our regulatory approach should no doubt be different from theirs, but the status quo—whereby the government heavily regulates who can access the most advanced US models, but most people can buy unlimited numbers of US AI chips and use them for any purpose—doesn't make much strategic sense.
(3) This should increase pressure on the US government to also fold Chinese models used/hosted in the United States into its own AI regulatory framework. It doesn't make sense for the US government to conduct pre-release safety testing of US companies' AI models, while deferring regulation of Chinese models used in the United States exclusively to the Chinese government—yet this will be the state of play when these regulations go into effect.
1
0
26
CFR's @ChrisRMcGuire reveals how Chinese companies can legally buy AI chips in the US and rent the compute back to China:
"Right now, anyone can make a company in Singapore and buy unlimited chips for a data center they're constructing in Malaysia, have a gigawatt of compute o Chinese companies, and all of that is totally legal. It’s bonkers."
"It's even worse. Chinese companies can buy AI chips in the United States and then rent it back to themselves in China, completely legal. Alibaba could set up a data center in the United States and rent the compute back to China."
"There’s obvious things we have to do to make these controls effective, and if we do, I think you really expand the lead."
"Right now, anyone can make a company in Singapore and buy unlimited chips for a data center they're constructing in Malaysia, have a gigawatt of compute o Chinese companies, and all of that is totally legal. It’s bonkers."
"It's even worse. Chinese companies can buy AI chips in the United States and then rent it back to themselves in China, completely legal. Alibaba could set up a data center in the United States and rent the compute back to China."
"There’s obvious things we have to do to make these controls effective, and if we do, I think you really expand the lead."
Three quick reactions to how the new Chinese export controls on AI, which give the Chinese government the authority to regulate the export, licensing, or publishing of any AI model (including open-source models), will impact the US government and industry:
(1) This introduces significant uncertainty for US businesses that host or rely on Chinese models. Particularly given that Chinese AI labs are moving toward licensing the use of open-source models to generate revenue, the Chinese government stating that it has the ability to curtail those licensing agreements at any time should make US businesses more hesitant to rely on those models going forward.
(2) This highlights why the US government also needs to adopt a framework that comprehensively governs the use of US AI technology overseas. The Chinese government clearly sees it as a national security imperative to understand who is utilizing or creating Chinese AI tech, and that it has the ability to regulate it--to include not only AI models, but also chips and data centers (even if China isn't exporting many of the latter). Our regulatory approach should no doubt be different from theirs, but the status quo—whereby the government heavily regulates who can access the most advanced US models, but most people can buy unlimited numbers of US AI chips and use them for any purpose—doesn't make much strategic sense.
(3) This should increase pressure on the US government to also fold Chinese models used/hosted in the United States into its own AI regulatory framework. It doesn't make sense for the US government to conduct pre-release safety testing of US companies' AI models, while deferring regulation of Chinese models used in the United States exclusively to the Chinese government—yet this will be the state of play when these regulations go into effect.
1
1
20
Council on Foreign Relations' @ChrisRMcGuire argues closing chip loopholes could turn America’s 6-month AI lead over China into 2 years:
"We’re in such a compute-constrained environment, no one can get enough compute, even domestically."
"JPMorgan had a study that said the na increase every year out to almost three hundred gigawatts by 2031. Everything that we send to China takes away from us."
"It’s the difference between us having a 25 to 1 compute advantage if they were using just their own chips, and probably 2 to 1 if they’re exploiting our controls. That’s the difference between a 6-month lead and a 2-year lead."
"We’re in such a compute-constrained environment, no one can get enough compute, even domestically."
"JPMorgan had a study that said the na increase every year out to almost three hundred gigawatts by 2031. Everything that we send to China takes away from us."
"It’s the difference between us having a 25 to 1 compute advantage if they were using just their own chips, and probably 2 to 1 if they’re exploiting our controls. That’s the difference between a 6-month lead and a 2-year lead."
Three quick reactions to how the new Chinese export controls on AI, which give the Chinese government the authority to regulate the export, licensing, or publishing of any AI model (including open-source models), will impact the US government and industry:
(1) This introduces significant uncertainty for US businesses that host or rely on Chinese models. Particularly given that Chinese AI labs are moving toward licensing the use of open-source models to generate revenue, the Chinese government stating that it has the ability to curtail those licensing agreements at any time should make US businesses more hesitant to rely on those models going forward.
(2) This highlights why the US government also needs to adopt a framework that comprehensively governs the use of US AI technology overseas. The Chinese government clearly sees it as a national security imperative to understand who is utilizing or creating Chinese AI tech, and that it has the ability to regulate it--to include not only AI models, but also chips and data centers (even if China isn't exporting many of the latter). Our regulatory approach should no doubt be different from theirs, but the status quo—whereby the government heavily regulates who can access the most advanced US models, but most people can buy unlimited numbers of US AI chips and use them for any purpose—doesn't make much strategic sense.
(3) This should increase pressure on the US government to also fold Chinese models used/hosted in the United States into its own AI regulatory framework. It doesn't make sense for the US government to conduct pre-release safety testing of US companies' AI models, while deferring regulation of Chinese models used in the United States exclusively to the Chinese government—yet this will be the state of play when these regulations go into effect.
2
3
26
SITUATION EXPLAINED: As AI-found bugs set off a wave of cyberattacks, Anthropic’s deputy security chief says "we tried to warn the world."
• Anthropic's deputy CISO Jason Clinton says things will get "a lot worse before they get better," and urged security leaders to ask their e basics, not AI
• This week, AI-found PHP flaws that could affect about 70% of websites sent government sites offline around the world
• Cloudflare’s CEO said exploit attempts jumped from thousands to millions within hours of the disclosure
• Clinton was replying to a critic who said AI labs "created this monster" and now want help cleaning it up
• Anthropic's view is that strong cyber models were coming anyway, so it built them first to patch critical software before attackers could
• Its Project Glasswing gives Claude Mythos Preview to partners like Apple, Google, and Microsoft to find and fix flaws
@theojaffee: "If you were Anthropic, what would you do differently? Would you just, like, not build cyber-capable models and then wait for some lab in China to do it and hack everyone in the country?"
• Anthropic's deputy CISO Jason Clinton says things will get "a lot worse before they get better," and urged security leaders to ask their e basics, not AI
• This week, AI-found PHP flaws that could affect about 70% of websites sent government sites offline around the world
• Cloudflare’s CEO said exploit attempts jumped from thousands to millions within hours of the disclosure
• Clinton was replying to a critic who said AI labs "created this monster" and now want help cleaning it up
• Anthropic's view is that strong cyber models were coming anyway, so it built them first to patch critical software before attackers could
• Its Project Glasswing gives Claude Mythos Preview to partners like Apple, Google, and Microsoft to find and fix flaws
@theojaffee: "If you were Anthropic, what would you do differently? Would you just, like, not build cyber-capable models and then wait for some lab in China to do it and hack everyone in the country?"
as bad as things have gotten in the past 4 days, they are going to get a lot worse before they get better. security leaders, never let a crisis go to waste: please go to your boards and ask for help right now. not for ai, for the basics. (we need to hold on for ~18 months.)
3
8
68
Xona CEO Brian Manning explains why losing GPS would disrupt financial markets, power grids, and phone networks, even if nobody needed directions:
"All of our financial infrastructure, all of our power grids, all of our telecom infrastructure use GPS for timing."
"When you make k and Japan, the way those are timestamped is using GPS. When your cellphone hands off from one tower to the next, it’s all using GPS timing."
"As you start to see these timing disruptions, a world without GPS is kicked back into the dark ages pretty quick."
@XonaSpace
"All of our financial infrastructure, all of our power grids, all of our telecom infrastructure use GPS for timing."
"When you make k and Japan, the way those are timestamped is using GPS. When your cellphone hands off from one tower to the next, it’s all using GPS timing."
"As you start to see these timing disruptions, a world without GPS is kicked back into the dark ages pretty quick."
@XonaSpace
1
1
33
Xona CEO Brian Manning explains how he’s using the SpaceX playbook to build a GPS alternative once reserved for superpowers:
"SpaceX is famous for the SpaceX algorithm, the process by which you design and build things. Step one is get rid of dumb requirements. All the same steps d over into Xona."
"Historically, GPS has been the realm of exclusively superpower governments."
"The only entities that have ever built a satellite navigation system operating on a global scale are the US government, the European government, the Chinese government, the Russian government, and now Xona."
@XonaSpace
"SpaceX is famous for the SpaceX algorithm, the process by which you design and build things. Step one is get rid of dumb requirements. All the same steps d over into Xona."
"Historically, GPS has been the realm of exclusively superpower governments."
"The only entities that have ever built a satellite navigation system operating on a global scale are the US government, the European government, the Chinese government, the Russian government, and now Xona."
@XonaSpace
$250M in new funding. $2.45B valuation. Six satellites launched this week.
Xona's Series D builds on major milestones in technology readiness, manufacturing scale, and commercial deployment as we accelerate toward serving users in 2027.
The next era of global navigation is closer than ever.
2
6
63
Relational Capital founder @MariaDavidson explains why the gap between frontier AI labs and construction sites is creating a massive opportunity for startups:
"The most important thing to me is always the founder. Do they have a unique insight that other people don’t have about market, about the way to win. Building a company has never been easier, and so it is actually so much more competitive than it was even 10 years ago."
"One of the areas that I love is investing in these really gritty, operationally heavy industries. As I look at SF and the insane progress we’re making on the frontier lab side, and then I think about my experience of spending a lot of the last eight years on job sites, the gap between those two is huge."
"Closing that diffusion gap requires an enormous amount of new companies to really be operationally excellent."
@RelationalCap
"The most important thing to me is always the founder. Do they have a unique insight that other people don’t have about market, about the way to win. Building a company has never been easier, and so it is actually so much more competitive than it was even 10 years ago."
"One of the areas that I love is investing in these really gritty, operationally heavy industries. As I look at SF and the insane progress we’re making on the frontier lab side, and then I think about my experience of spending a lot of the last eight years on job sites, the gap between those two is huge."
"Closing that diffusion gap requires an enormous amount of new companies to really be operationally excellent."
@RelationalCap
I'm excited to launch Relational Capital, a $50m solo GP fund.
I spent the last 8 years as a founder, building @usekojo into the largest US construction procurement platform.
Now I'm backing founders at the earliest stages with $1m-$5m checks.
An abundant future lies ahead 🇺🇸
7
1
43
Boom Supersonic CEO @bscholl explains why we got Star Trek's iPads but not its warp drives:
"Star Trek Next Generation, touchscreens, everyone’s walking around with an iPad. We got that, but we didn’t get the photon torpedoes, we didn’t get the warp drive. We got all the low idn’t get the high energy sci-fi."
"Thomas Edison flipped the switch on the very first electric grid in 1882. From 1882 through to 1973, there was exponential increase in power production per capita, doubling about every 11 years. And then in 1973, something went wrong, and that trend stopped."
"Since the year 2000, our population has increased faster than our energy. The amount of energy available to the average American is actually less now than it was 26 years ago."
"We’ve shipped all the manufacturing to Asia. I think we all know now that was a terrible idea. China’s got three times the energy production of the US."
"The way to bring it back is not bring it back, but invent the next generation of manufacturing technology here. We’re gonna need a lot of power for that. And of course, AI."
@boomsupersonic
"Star Trek Next Generation, touchscreens, everyone’s walking around with an iPad. We got that, but we didn’t get the photon torpedoes, we didn’t get the warp drive. We got all the low idn’t get the high energy sci-fi."
"Thomas Edison flipped the switch on the very first electric grid in 1882. From 1882 through to 1973, there was exponential increase in power production per capita, doubling about every 11 years. And then in 1973, something went wrong, and that trend stopped."
"Since the year 2000, our population has increased faster than our energy. The amount of energy available to the average American is actually less now than it was 26 years ago."
"We’ve shipped all the manufacturing to Asia. I think we all know now that was a terrible idea. China’s got three times the energy production of the US."
"The way to bring it back is not bring it back, but invent the next generation of manufacturing technology here. We’re gonna need a lot of power for that. And of course, AI."
@boomsupersonic
Today Boom unveils our first completed supersonic turbine core, a 2GW Superfactory, and—one more thing—the world's most advanced all-digital foundry to cast single crystal turbine blades and vanes x.com/i/article/2113…
6
11
118
Boom Supersonic CEO @bscholl reveals why most data center water usage comes from cooling turbines, not servers, and how his new turbine doesn't need a single drop:
"Number one is you just can’t get power from the grid, so now you gotta get turbines. Can’t get turbines without ying exorbitant prices on the secondary market."
"The public’s got concerns. If you’re gonna pull electricity from the grid, our electricity price is gonna go up. What about water usage? The dirty little secret is that most data center water usage isn’t the server cooling, it’s actually the turbine cooling. This turbine doesn’t need a drop of water, and it lets you make your own electricity."
"The third thing is noise. We’re gonna make this an ultra quiet turbine. It’s from the same team that made sonic boom go away entirely, so I’m pretty sure we can make a turbine quiet too."
@boomsupersonic
"Number one is you just can’t get power from the grid, so now you gotta get turbines. Can’t get turbines without ying exorbitant prices on the secondary market."
"The public’s got concerns. If you’re gonna pull electricity from the grid, our electricity price is gonna go up. What about water usage? The dirty little secret is that most data center water usage isn’t the server cooling, it’s actually the turbine cooling. This turbine doesn’t need a drop of water, and it lets you make your own electricity."
"The third thing is noise. We’re gonna make this an ultra quiet turbine. It’s from the same team that made sonic boom go away entirely, so I’m pretty sure we can make a turbine quiet too."
@boomsupersonic
Today Boom unveils our first completed supersonic turbine core, a 2GW Superfactory, and—one more thing—the world's most advanced all-digital foundry to cast single crystal turbine blades and vanes x.com/i/article/2113…
3
8
118
Boom Supersonic CEO @bscholl explains why the race to AI superintelligence could come down to turbine blades that only a couple of companies on Earth can manufacture:
"The hardest stuff are the turbine blades that spin at ten thousand RPM in an environment that is about three r than Starship on re-entry."
"They have to be actively cooled, or they will melt. @elonmusk talked about this a couple days ago on the SpaceX earnings call."
"Those blades are the bottleneck in all of turbines, and turbines are the bottleneck in all of energy, and energy is the bottleneck in superintelligence."
"You take this molten metal, and it freezes in such a way that the atoms all line up in one perfect grid. Billions and trillions of atoms in one perfect grid. There are only a couple companies on the planet that can do this, they’re supply-constrained, and they’re certainly not gonna hand their manufacturing capacity to a startup."
"We’re gonna be doing that ourselves. We are building an all-digital foundry. Our goal is to go in twenty-four hours from a turbine blade design to the world’s most advanced single crystal casting. I don’t think there’s anywhere you can actually do that today."
@boomsupersonic
"The hardest stuff are the turbine blades that spin at ten thousand RPM in an environment that is about three r than Starship on re-entry."
"They have to be actively cooled, or they will melt. @elonmusk talked about this a couple days ago on the SpaceX earnings call."
"Those blades are the bottleneck in all of turbines, and turbines are the bottleneck in all of energy, and energy is the bottleneck in superintelligence."
"You take this molten metal, and it freezes in such a way that the atoms all line up in one perfect grid. Billions and trillions of atoms in one perfect grid. There are only a couple companies on the planet that can do this, they’re supply-constrained, and they’re certainly not gonna hand their manufacturing capacity to a startup."
"We’re gonna be doing that ourselves. We are building an all-digital foundry. Our goal is to go in twenty-four hours from a turbine blade design to the world’s most advanced single crystal casting. I don’t think there’s anywhere you can actually do that today."
@boomsupersonic
Today Boom unveils our first completed supersonic turbine core, a 2GW Superfactory, and—one more thing—the world's most advanced all-digital foundry to cast single crystal turbine blades and vanes x.com/i/article/2113…
7
5
52
Boom Supersonic CEO @bscholl reveals how building the XB-1 supersonic aircraft exposed a broken aerospace supply chain where parts took 18 months to arrive despite taking just 3 days to manufacture:
"This engine is built by the same team that built our XB-1 supersonic aircraft, the sound barrier last year, proved that sonic boom didn’t have to be a thing, and resulted in legalization of supersonic flight. We probably learned ten times more than we thought we would."
"The traditional aerospace supply chain in this country, I have no polite words for it. Let’s just say it sucks. Our engineers spent more time waiting for parts than they did designing, assembling, and building."
"We coined a term we call the slacker index. How long it takes to get something divided by how long it actually takes to make it. There were things that would take 3 days to manufacture but take 18 months to get."
"We decided it was time to go vertically integrated, not just designing the engine ourselves, but actually making the parts too."
@boomsupersonic
"This engine is built by the same team that built our XB-1 supersonic aircraft, the sound barrier last year, proved that sonic boom didn’t have to be a thing, and resulted in legalization of supersonic flight. We probably learned ten times more than we thought we would."
"The traditional aerospace supply chain in this country, I have no polite words for it. Let’s just say it sucks. Our engineers spent more time waiting for parts than they did designing, assembling, and building."
"We coined a term we call the slacker index. How long it takes to get something divided by how long it actually takes to make it. There were things that would take 3 days to manufacture but take 18 months to get."
"We decided it was time to go vertically integrated, not just designing the engine ourselves, but actually making the parts too."
@boomsupersonic
Boom! We cracked it! Today we are introducing Boomless Cruise—supersonic flights up to 50% faster with no audible sonic boom.
We quietly (har har) demo'd this on XB-1's first supersonic flight—three times actually. 🧵👇
1
3
27
Ex-OpenAI science head @kevinweil warns America risks losing its pharmaceutical capabilities to China the same way it lost manufacturing:
"We started by offshoring the low-skilled manufacturing work. Then over time we pushed more and more out of the country, and we lost the amazing quote from Andy Grove: abandoning today’s commodity manufacturing can lock you out of tomorrow’s emerging industry. That is the path that we are on with pharma. It’s not too late to take it back."
"If you’re gonna bring it back, you’ve got to change the economics. There’s an economic reason why this all moved out. The way that you change the economics is with robotics and AI and automation."
@warpsciences
"We started by offshoring the low-skilled manufacturing work. Then over time we pushed more and more out of the country, and we lost the amazing quote from Andy Grove: abandoning today’s commodity manufacturing can lock you out of tomorrow’s emerging industry. That is the path that we are on with pharma. It’s not too late to take it back."
"If you’re gonna bring it back, you’ve got to change the economics. There’s an economic reason why this all moved out. The way that you change the economics is with robotics and AI and automation."
@warpsciences
On #2: Andy Grove once said “abandoning today's 'commodity' manufacturing can lock you out of tomorrow's emerging industry” and nowhere is this more true than in pharma and life sciences.
Over the last 20 years we have shifted much of the drug discovery process to large Chinese contract research organizations (CROs) like Wuxi—a company which was added to the DoW’s list of “Chinese military companies” in June.
We’ve seen this multiple times now. When the work moves overseas, the expertise and infrastructure follow. We're at risk of losing the ability to turn our ideas into medicines. That’s an unacceptable dependency for something so central to our health and our future.
0
5
71
Ex-OpenAI science head @kevinweil explains why AI is making breakthroughs in math and coding but still struggles to discover new medicines:
"We now live in a world where it doesn’t matter whether you learned how to program, you can turn any idea you have into an actual working s."
"You got models solving Navier-Stokes and making progress on the Riemann hypothesis. But I think the biggest impact that AI will have for the world will be in sciences."
"These fields of science are so empirical. You don’t just think really hard about how you find a molecule that binds to the protein. You run experiments. If we want models to do that, the models need to see that data at scale."
"This is why there’s such a difference between model performance in mathematics and software engineering and model performance in science."
@warpsciences
"We now live in a world where it doesn’t matter whether you learned how to program, you can turn any idea you have into an actual working s."
"You got models solving Navier-Stokes and making progress on the Riemann hypothesis. But I think the biggest impact that AI will have for the world will be in sciences."
"These fields of science are so empirical. You don’t just think really hard about how you find a molecule that binds to the protein. You run experiments. If we want models to do that, the models need to see that data at scale."
"This is why there’s such a difference between model performance in mathematics and software engineering and model performance in science."
@warpsciences
On #1: I was previously CPO and VP of Science at @OpenAI, and I couldn’t be more optimistic about the future. Anyone can use AI today to build an idea in minutes, and we’ve just seen models solve Navier-Stokes and make progress on the Riemann Hypothesis.
But the reality is that AI is not yet at the same level in the physical sciences, especially in places like drug discovery where it could do the greatest good.
The reason is simple: these are highly emergent and empirical sciences, and they move forward through data and experiment. The models don’t have access to the type and volume of experimental data they need.
6
10
106
Ex-OpenAI science head @kevinweil reveals how America outsourced a critical 3-5-year stage of drug development entirely to China:
"Imagine you and I start a drug discovery company. You have a hypothesis that if we could inhibit this protein with a drug, we have a positive effect You find a molecule that binds to the protein, but most likely that first molecule is gonna be toxic, or not soluble, or not very potent. Instead of a little pill, you have to take a kilogram of it."
"You spend the next three to five years making variations of this molecule hoping that you find one that is less toxic, more potent, more soluble."
"The thing that people don’t know is that whole three to five years of making and testing molecules, we’ve outsourced it entirely to China. We cannot depend on a geopolitical adversary to do this work for us."
@warpsciences
"Imagine you and I start a drug discovery company. You have a hypothesis that if we could inhibit this protein with a drug, we have a positive effect You find a molecule that binds to the protein, but most likely that first molecule is gonna be toxic, or not soluble, or not very potent. Instead of a little pill, you have to take a kilogram of it."
"You spend the next three to five years making variations of this molecule hoping that you find one that is less toxic, more potent, more soluble."
"The thing that people don’t know is that whole three to five years of making and testing molecules, we’ve outsourced it entirely to China. We cannot depend on a geopolitical adversary to do this work for us."
@warpsciences
💥 Introducing Warp Sciences, a new company applying AI and robotics to help the world develop new medicines faster.
We’ve raised a $215M Series A led by @lightspeedvp to bring frontier AI out of the browser and into the lab, and build scientific capacity here in America.
I co-founded Warp with my long-time friend @dpatil, FAIR co-founder @manohar_paluri, and drug hunter Sol Reisberg. The early team is full of incredible people like @bkspears9, who helped achieve Q>1 fusion at LLNL.
2
3
34
Ex-OpenAI science head @kevinweil on his $215M plan to build robotic labs that run 24/7 and bring drug research back from China:
"We’re building robotic labs that do science at scale, and we’re doing it here in America. Doing science at scale is the way we generate the training tely need."
"We need to reshore scientific capabilities from China so that we don’t lose our pharma industry the way that we lost solar and drones and manufacturing."
"We’re not developing drugs on our own, we are building the infrastructure so that everybody who does develop drugs can go faster."
@warpsciences
"We’re building robotic labs that do science at scale, and we’re doing it here in America. Doing science at scale is the way we generate the training tely need."
"We need to reshore scientific capabilities from China so that we don’t lose our pharma industry the way that we lost solar and drones and manufacturing."
"We’re not developing drugs on our own, we are building the infrastructure so that everybody who does develop drugs can go faster."
@warpsciences
💥 Introducing Warp Sciences, a new company applying AI and robotics to help the world develop new medicines faster.
We’ve raised a $215M Series A led by @lightspeedvp to bring frontier AI out of the browser and into the lab, and build scientific capacity here in America.
I co-founded Warp with my long-time friend @dpatil, FAIR co-founder @manohar_paluri, and drug hunter Sol Reisberg. The early team is full of incredible people like @bkspears9, who helped achieve Q>1 fusion at LLNL.
8
17
165
Project Eleven CEO @apruden08 warns AI will put everything exposed to the internet under constant attack, just as crypto has been for years:
"No one cares about security until they do. What people are coming to appreciate about security in the AI era is that everything that’s be under constant attack."
"The cost of attacking this infrastructure, if you’re a hacker, has dropped to basically zero. You can spin up any amount of agents and start pressure testing infrastructure all the way down to open source libraries."
"This is similar to how crypto has been from the beginning. I’m trying to think of a word that’s nicer than paranoid, but you kind of need to be paranoid. In the post-AI cybersecurity landscape, everyone is going to need to have a security-first mindset."
@projecteleven
"No one cares about security until they do. What people are coming to appreciate about security in the AI era is that everything that’s be under constant attack."
"The cost of attacking this infrastructure, if you’re a hacker, has dropped to basically zero. You can spin up any amount of agents and start pressure testing infrastructure all the way down to open source libraries."
"This is similar to how crypto has been from the beginning. I’m trying to think of a word that’s nicer than paranoid, but you kind of need to be paranoid. In the post-AI cybersecurity landscape, everyone is going to need to have a security-first mindset."
@projecteleven
Assumes three things:
(1) that the state of the art models are widely available/open-source. That is contrary to what my expectations are w.r.t. incentives. Whether for safety or revenue, OpenAI and Anthropic have reasons to gate their best cyber models and make you pay up for an AI-audit basically.
That doesn't slow down sophisticated attackers much: nation-states will pay or build their own anyway. What it does is make AI-assisted defense something only large organizations can afford, leaving the long tail, where most of the easy targets are, exposed. (goes without saying that open source libraries maintained by volunteers are probably toast no matter what)
(2) that the security vulnerabilities themselves are "shallow" and easily detectable. If bugs are deep, finding them takes a lot of tokens per line of code, and that's where the asymmetry bites: because defenders have to secure the whole surface area of their codebase against attack, dollar for dollar they will not be able to spend as much in tokens to *harden* every LOC as the attackers will be able to spend to *break* one specific point.
A few reasons to doubt (2):
a. The PHP vulnerability was a very deep bug that many people overlooked.
b. It's been shown that just spending more tokens can get better results. (e.g. the Claude Riemann-Zeta result, where the engineer basically just kept prompting the model to "keep trying")
c. ecdsa dot fail, the more tokens get thrown at the problem, the more progress gets made. Implies that spending more on tokens surfaces more insights as you get deeper into the problem.
On their own, b. and c. cut both ways: if more tokens produce better results, defenders get more tokens too. They favor the attacker because of the asymmetry: the defender has to spend tokens hardening every line, while the attacker only has to spend them on one.
(3) that the pace of software development slows down to the pace of security engineering. If anything, AI is vastly *accelerating* the product/engineering cycles, and exponentially *more code* will be produced in the coming years. Hard to imagine that security tools will be able to keep pace (security is always ignored by everyone until it's impossible not to).
Defenders do have home-field advantages: source access, and the chance to scan code before it ships. But that only helps the orgs that actually do it, which brings us back to (1) and (3).
So it's not obvious to me that AI favors the defense. In fact, at the limit things will settle to roughly where they stand today: attacker has slight advantage based on the asymmetry of resources they can bring to bear against one, carefully selected target.
AI tools can certainly help improve security, but not enough to overcome this fundamental asymmetry and flip the advantage to the defender, imho.
7
11
41
SITUATION EXPLAINED: DeepSeek is closing a $15 billion raise, and its backers include Tencent, CATL, and a dishwasher maker.
• The round is about 100B yuan, double the original 50B yuan target, at roughly 500B yuan (about $70B) before the new money
• Carmaker Geely, delivery maker GigaDevice joined too
• Most of the backers are companies in China’s tech supply chain rather than VC firms, which could make DeepSeek a core part of that chain
• That valuation is a fraction of what top US labs are worth, with Anthropic’s last round valuing it at $965B
• DeepSeek hired CITIC Securities in September to prepare a possible IPO on Shanghai’s STAR Market
@theojaffee: "It’s a sort of flywheel where DeepSeek helps Huawei build chips that are then used to train the next DeepSeek model that then makes DeepSeek money that they can use to invest in chips made by Huawei. It’s a virtuous cycle."
• The round is about 100B yuan, double the original 50B yuan target, at roughly 500B yuan (about $70B) before the new money
• Carmaker Geely, delivery maker GigaDevice joined too
• Most of the backers are companies in China’s tech supply chain rather than VC firms, which could make DeepSeek a core part of that chain
• That valuation is a fraction of what top US labs are worth, with Anthropic’s last round valuing it at $965B
• DeepSeek hired CITIC Securities in September to prepare a possible IPO on Shanghai’s STAR Market
@theojaffee: "It’s a sort of flywheel where DeepSeek helps Huawei build chips that are then used to train the next DeepSeek model that then makes DeepSeek money that they can use to invest in chips made by Huawei. It’s a virtuous cycle."
4
5
64
SITUATION EXPLAINED: Elon Musk just backed a $3.5 trillion merger of SpaceX and Tesla, which would be the biggest deal in history.
• Musk backed it on an investor call, but he’s conflicted and has to stay out of the talks, so insiders worry it’ll draw shareholder scrutiny and dependent committees at both companies are weighing SpaceX buying Tesla or a new holding company, likely in an all-stock deal
• It echoes Tesla’s 2016 SolarCity buyout, where Musk was also on both sides, which led to years of lawsuits
• Tesla shareholders would likely want a premium, which Wedbush’s Dan Ives puts at about $500–550 a share
• Musk has said since 2024 he wants about 25% voting control of Tesla, and analysts say a merger would give him control of the combined company
• The companies’ AI and chip work now overlaps, and closing before Trump’s term ends could help the deal avoid a tougher antitrust review
@theojaffee: "Elon has just been grinding away at the exponential for years and years and years."
• Musk backed it on an investor call, but he’s conflicted and has to stay out of the talks, so insiders worry it’ll draw shareholder scrutiny and dependent committees at both companies are weighing SpaceX buying Tesla or a new holding company, likely in an all-stock deal
• It echoes Tesla’s 2016 SolarCity buyout, where Musk was also on both sides, which led to years of lawsuits
• Tesla shareholders would likely want a premium, which Wedbush’s Dan Ives puts at about $500–550 a share
• Musk has said since 2024 he wants about 25% voting control of Tesla, and analysts say a merger would give him control of the combined company
• The companies’ AI and chip work now overlaps, and closing before Trump’s term ends could help the deal avoid a tougher antitrust review
@theojaffee: "Elon has just been grinding away at the exponential for years and years and years."
2
1
66
Kong CEO @sonicaghi warns AI is giving hackers nation-state-level capabilities while companies deploy agents they can't secure:
"In the last 24 hours we had PHP2Shell, a bunch of gateways got compromised, and the frequency’s increasing. You have now nationwide type of attacker dels, and the companies haven’t defended themselves through security models."
"Agentic enterprise and agentic coding is creating a whole new surface that most of the time is non-deterministic. That creates a lot of empty spots to be attacked."
"We do rate limiting, blocking, and a kill switch. This is becoming a large risk into boardrooms of the Fortune 500s."
@kong
"In the last 24 hours we had PHP2Shell, a bunch of gateways got compromised, and the frequency’s increasing. You have now nationwide type of attacker dels, and the companies haven’t defended themselves through security models."
"Agentic enterprise and agentic coding is creating a whole new surface that most of the time is non-deterministic. That creates a lot of empty spots to be attacked."
"We do rate limiting, blocking, and a kill switch. This is becoming a large risk into boardrooms of the Fortune 500s."
@kong
Two critical vulnerabilities have been identified in PHP, a widely used programming language that powers websites and web applications. These vulnerabilities (CVE-2026-109618 and CVE-2026-109619) could allow malicious cyber actors to remotely execute code in PHP core, potentially data and networks.
The FBI urges organizations to take the following steps now:
➡️ If an immediate update is not feasible it is recommended that organizations take the system offline until it can be fully mitigated.
➡️ Immediately update affected PHP installations following the guidance on php.net. Inventory where PHP runs in your environment, prioritizing internet-facing systems and vendor products that bundle PHP.
➡️ Monitor web servers for suspicious activity, including PHP processes unexpectedly launching command shells or downloading files. Ensure appropriate process-monitoring capabilities are in place to help identify potential exploitation.
➡️ If you find signs of compromise, isolate affected systems, preserve your web server and PHP logs, and report it to your local FBI field office or ic3.gov. Those reports help us identify the actors involved and warn other organizations before they become victims.
For the latest information, follow PHP’s announcement on php.net, CERT-EU Security Advisory 2026-018, and CISA’s Known Exploited Vulnerabilities catalog.
2
2
17
Kong CEO @sonicaghi reveals token traffic through its gateways is growing 140x year over year, approaching 1 quadrillion tokens:
"The number of tokens is growing exponentially. For the subset that we see, we are running about 500 trillion tokens, going to a quadrillion soon, 0x year over year."
"The cost per token is going down, and inference will still be exploding. We’re gonna live in this enterprise world that is more multi-model, cheaper tokens, cheaper levels of intelligence."
"The next pieces are, how do we secure it? How do we govern it? How do we cost control it?"
@kong
"The number of tokens is growing exponentially. For the subset that we see, we are running about 500 trillion tokens, going to a quadrillion soon, 0x year over year."
"The cost per token is going down, and inference will still be exploding. We’re gonna live in this enterprise world that is more multi-model, cheaper tokens, cheaper levels of intelligence."
"The next pieces are, how do we secure it? How do we govern it? How do we cost control it?"
@kong
3
4
25
Former DOE security engineer @asynchronous_x explains why America's critical infrastructure hasn't been attacked more often despite its networks being riddled with security holes:
"The work I used to do for the Department of Energy was around critical infrastructure. They’ve u don’t know that they’re there, so why would anyone attack us. As soon as you took any sort of magnifying glass to their networks, they were usually riddled with holes."
"The reason they haven’t gotten attacked so much is because the repercussion is so much higher. If you compromise a water treatment facility and you accidentally end up killing people, that’s gonna have extreme repercussions versus if you steal some cryptocurrency."
"We’ve been relying on political and other methods for keeping them safe rather than actually making them secure."
"The work I used to do for the Department of Energy was around critical infrastructure. They’ve u don’t know that they’re there, so why would anyone attack us. As soon as you took any sort of magnifying glass to their networks, they were usually riddled with holes."
"The reason they haven’t gotten attacked so much is because the repercussion is so much higher. If you compromise a water treatment facility and you accidentally end up killing people, that’s gonna have extreme repercussions versus if you steal some cryptocurrency."
"We’ve been relying on political and other methods for keeping them safe rather than actually making them secure."
Two critical vulnerabilities have been identified in PHP, a widely used programming language that powers websites and web applications. These vulnerabilities (CVE-2026-109618 and CVE-2026-109619) could allow malicious cyber actors to remotely execute code in PHP core, potentially data and networks.
The FBI urges organizations to take the following steps now:
➡️ If an immediate update is not feasible it is recommended that organizations take the system offline until it can be fully mitigated.
➡️ Immediately update affected PHP installations following the guidance on php.net. Inventory where PHP runs in your environment, prioritizing internet-facing systems and vendor products that bundle PHP.
➡️ Monitor web servers for suspicious activity, including PHP processes unexpectedly launching command shells or downloading files. Ensure appropriate process-monitoring capabilities are in place to help identify potential exploitation.
➡️ If you find signs of compromise, isolate affected systems, preserve your web server and PHP logs, and report it to your local FBI field office or ic3.gov. Those reports help us identify the actors involved and warn other organizations before they become victims.
For the latest information, follow PHP’s announcement on php.net, CERT-EU Security Advisory 2026-018, and CISA’s Known Exploited Vulnerabilities catalog.
2
19
65
Former DOE security engineer @asynchronous_x warns AI is narrowing the gap between junior hackers and agencies like the NSA:
"We used to have this thing in industry where we would say, this is pretty secure, unless your threat actor is the NSA."
"Now with AI tools you can take less expertise, and they can make more sophisticated attacks. Somebody that would’ve been a less high-level threat actor before could potentially become a much more dangerous threat actor."
"There’s plenty in the industry that think we should limit access to these tools because it can make smaller threat actors much more dangerous."
"We used to have this thing in industry where we would say, this is pretty secure, unless your threat actor is the NSA."
"Now with AI tools you can take less expertise, and they can make more sophisticated attacks. Somebody that would’ve been a less high-level threat actor before could potentially become a much more dangerous threat actor."
"There’s plenty in the industry that think we should limit access to these tools because it can make smaller threat actors much more dangerous."
Some stats on #php2shell. In the 10 hours before it was publicly announced, across the @Cloudflare network we saw thousands of attempts to exploit it. That jumped to tens of thousands in the 45 minutes post announcement, hundreds of thousands within 2 hours, and millions now.
2
9
29
Former DOE security engineer @asynchronous_x explains how a miscommunication turned a planned PHP2Shell fix into a race against hackers:
"It should have been the best-case scenario. Security researchers identified these vulnerabilities, developed proof of concept exploits in ming the PHP core team for fixes. That whole part was working fine."
"The reason this one’s unique is there was basically a leak of the vulnerability before the disclosure. That was supposed to be released on October 22nd, today. Unfortunately, there was a miscommunication, and they released some of the details early. So there was this panic race where the problem is out in the wild, but we don’t have a fix prepared yet."
"The thing it’s taught the security realm is that we probably can’t rely on coordinated disclosure anymore. If the speed at which people can develop exploits is now so fast that we can’t possibly rely on things not getting leaked, then it’s not something we can lean on as much as we used to."
"It should have been the best-case scenario. Security researchers identified these vulnerabilities, developed proof of concept exploits in ming the PHP core team for fixes. That whole part was working fine."
"The reason this one’s unique is there was basically a leak of the vulnerability before the disclosure. That was supposed to be released on October 22nd, today. Unfortunately, there was a miscommunication, and they released some of the details early. So there was this panic race where the problem is out in the wild, but we don’t have a fix prepared yet."
"The thing it’s taught the security realm is that we probably can’t rely on coordinated disclosure anymore. If the speed at which people can develop exploits is now so fast that we can’t possibly rely on things not getting leaked, then it’s not something we can lean on as much as we used to."
OpenAI models helped researchers—operating as part of our Patch the Planet initiative—discover critical security flaws in PHP, an open-source programming language used to run an estimated 70% of websites.
Over the past two months, we led a coalition of maintainers, researchers, and security providers to prepare fixes and protections ahead of disclosure. This duration was necessary because the fix was quite intricate and required a major change to the PHP engine. Every site using PHP can now safely upgrade their version to stay secure, in addition to applying mitigations from various security vendors.
Due to the severity of these flaws and PHP’s widespread use, we consider this bug to be one of the most impactful vulnerabilities in recent memory. It also marks a new phase, where AI can find security bugs with internet-scale impact.
This is an example of the Defender’s Window in action: to our knowledge, frontier AI was able to find this vulnerability ahead of attackers, giving defenders time to prepare.
openai.com/index/coordina…Community noteOpenAI may claim a role in the coordinated response and patching process. But the initial discovery and exploit development, however, occurred several months earlier by an independent researcher.
x.com/lcfr_hax/statu…
3
2
24
Synthesia CEO @vriparbelli reveals how LLMs have become surprisingly good at video editing, making it possible to create entire videos through chat:
"You have to write a script. LLMs are pretty good at that. You have to generate the actual assets, that’s usually what you capture hen there’s a process after that where someone takes those assets, takes the script, and edits everything together."
"The big thing that’s happened in the last couple of months is that LLMs have now become an amazing video editor as well. All the motion graphics, text on screen, how to delineate the different scenes."
"I don’t think anyone really expected we’d get this good this fast. We’ve gotten very, very close to what it would be like to work with a really good video editor."
@synthesiaIO
"You have to write a script. LLMs are pretty good at that. You have to generate the actual assets, that’s usually what you capture hen there’s a process after that where someone takes those assets, takes the script, and edits everything together."
"The big thing that’s happened in the last couple of months is that LLMs have now become an amazing video editor as well. All the motion graphics, text on screen, how to delineate the different scenes."
"I don’t think anyone really expected we’d get this good this fast. We’ve gotten very, very close to what it would be like to work with a really good video editor."
@synthesiaIO
Syren Video is now live and FREE to try!
ChatGPT moment for agentic video.
Prompt → agency-quality AI video → chat to edit.
Powered by Opus 5.5 + Syren renderer w/ 3D support.
Browser or Claude MCP.
Insane examples in thread 🤯🤯🤯
Try it free here: shorturl.at/r57qy
2
3
23
.@deredleritt3r explains why OpenAI's latest math breakthroughs have him convinced AI could be on the brink of RSI:
"When I used GPT-6 Astra, I messed around with a couple of ciphers that were previously unresolved, from World War I, and I successfully had Astra decrypt a couple ne of the results was the model having the key, and brute forcing the solution by searching through the entire space of possibilities. I have not been able to get Astra to come up with the key on its own based on some novel idea."
"The math results recently achieved by OpenAI’s internal model have really updated me. I think Terence Tao said about the quasi-Riemann solution that it was a combination of three or four existing ideas but arranged in a very unusual way."
"I’m guessing an internal model at OpenAI is probably still not able to achieve RSI. But it feels like we’re really not very far away. Once the models are able to do something, in 12 months or sooner, they’ll be really, really good at it. I think we’re on the brink."
"When I used GPT-6 Astra, I messed around with a couple of ciphers that were previously unresolved, from World War I, and I successfully had Astra decrypt a couple ne of the results was the model having the key, and brute forcing the solution by searching through the entire space of possibilities. I have not been able to get Astra to come up with the key on its own based on some novel idea."
"The math results recently achieved by OpenAI’s internal model have really updated me. I think Terence Tao said about the quasi-Riemann solution that it was a combination of three or four existing ideas but arranged in a very unusual way."
"I’m guessing an internal model at OpenAI is probably still not able to achieve RSI. But it feels like we’re really not very far away. Once the models are able to do something, in 12 months or sooner, they’ll be really, really good at it. I think we’re on the brink."
We’re releasing a broad range of new mathematical results produced by an internal frontier model.
We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and public recommendations to inform how we release these results.
github.com/openai/math
2
4
117
.@deredleritt3r warns the emerging cyber crisis feels like February 2020, when nobody knew how bad COVID would get:
"This reminds me of February 2020, when we heard about COVID-19 for the first time, and everyone was like, how bad is it gonna be?"
"The US federal government, ess to Mythos since basically February of this year, I am hoping and praying that they have, secretly from the public, strengthened a lot of the really critical systems. But we can’t know for sure."
"Right now anything that is vulnerable is going to be attacked. Two-factor authentication, change your passwords. Do as much as you can."
"This reminds me of February 2020, when we heard about COVID-19 for the first time, and everyone was like, how bad is it gonna be?"
"The US federal government, ess to Mythos since basically February of this year, I am hoping and praying that they have, secretly from the public, strengthened a lot of the really critical systems. But we can’t know for sure."
"Right now anything that is vulnerable is going to be attacked. Two-factor authentication, change your passwords. Do as much as you can."
Just in case we lose our internets forever over the next few hours, let this be my final post, thank you and godspeed
3
6
45
.@deredleritt3r argues real RSI begins when AI can make Ilya Sutskever-level breakthroughs on its own:
"It’s a good thing that you’re talking to a lawyer because definitions is what we do. I focus on the word self. You need to be able to train a model that is your successor and an you are. Until we have that, I don’t think it’s RSI."
"Right now we don’t have RSI, and I don’t think that the labs have it internally either."
"We will have it when a model is capable of fully autonomously coming up with really creative ways to make the new model more capable. Ilya Sutskever type brilliant moments. I think that’s coming pretty soon."
"It’s a good thing that you’re talking to a lawyer because definitions is what we do. I focus on the word self. You need to be able to train a model that is your successor and an you are. Until we have that, I don’t think it’s RSI."
"Right now we don’t have RSI, and I don’t think that the labs have it internally either."
"We will have it when a model is capable of fully autonomously coming up with really creative ways to make the new model more capable. Ilya Sutskever type brilliant moments. I think that’s coming pretty soon."
5
5
43
.@deredleritt3r warns the cyber apocalypse may have already begun:
"When you have an event followed within 30 minutes by another event, followed in an hour by another event, and so on every couple of hours, it starts feeling like we’ve crossed some kind of threshold."
"This has I wouldn’t be surprised if it’s happening now. It certainly seems to be happening now."
"When you have an event followed within 30 minutes by another event, followed in an hour by another event, and so on every couple of hours, it starts feeling like we’ve crossed some kind of threshold."
"This has I wouldn’t be surprised if it’s happening now. It certainly seems to be happening now."
Just in case we lose our internets forever over the next few hours, let this be my final post, thank you and godspeed
3
22
127