Crazy 72 hours in security. Three major incidents dropped back-to-back: eMudhra, PingCAP (TiDB Cloud), and PHP. Here is the short version of what happened and who got hit. 1. eMudhra (Trusted CA compromise) eMudhra generated 8 intermediate CA certificates in April 2025 (cross-signed by Sectigo). A bug in their HSM key-wrapping test script created a second set of P-384 keys where the private keys were single-byte numbers under 256. Operators picked the wrong key labels and signed them into real CA certs. One of the intermediate CA private keys was literally the number 4. The other seven: 56, 66, 115, 155, 156, 230, 237. Three weeks later (May 2025), their CRL signatures started failing because the CRL signer was using the real keys instead of the weak ones. Instead of asking why the keys didn't match, someone switched the CRL signer to the weak keys and closed the ticket. • Browsers are safe: Chrome, Firefox, Apple, and Microsoft already blocked them, and zero rogue certs hit CT logs. • Servers are not: curl, Python (certifi), Go binaries, and Docker base images don't check CRLs. Update ca-certificates, certifi, and Go (1.27.3 / 1.26.10). • 10-second check on any server: curl dv-g3b-emudhra.we-have-a-pro… (TLS error = safe. 200 OK = still vulnerable.) 2. PingCAP / TiDB Cloud (10.6 TB breach) Attackers found GCP and Azure credentials committed inside PingCAP's own repos. They set up a replication feed of TiDB Cloud's production control plane and exported the entire database every 6 hours for 22 straight days. 88 full exports, zero alerts. 200 GB is already out on torrents. Beyond the 413 private repos and 20.6M customer SQL queries logged from their web SQL Editor, PingCAP engineers were pasting customer secrets into Jira tickets. Who is affected in the leaked records: • Live credentials exposed in support tickets: - Pinterest: AWS keys for their pingcap-logs S3 bucket - CardX / Silverlake (NPCI-linked env): AWS admin key in a backup log with full access to 258 external S3 buckets - DeBank: 38 presigned S3 URLs - An unnamed AI/dev platform: a 314 MB slow-query log attached to a ticket containing 8,027 live customer secrets (an AWS root key, a GitHub admin PAT over 1,701 repos, 7,485 AWS STS tokens, and 192 OpenAI/Anthropic/Gemini keys) • Tenants with infra metadata, VPC CIDRs, app-server IPs, or contract pricing exposed: - Tech: Atlassian, Databricks, Airbnb, Baidu, LINE, Nylas, Persona, PatSnap - E-comm & Travel: Booking.com, Shopee, Flipkart, Mercari, ZOZO, AEON, Delhivery, Klook - Gaming: Capcom, Square Enix, Konami, Bandai Namco, Cygames, Colopl, DMM - Finance: PayPay, SoftBank, SBI, SB Payment Service, Fujitsu, TOPPAN, Bank of China, China Life - Crypto: Bitget (said only public trade logs were on it and they already left), Pionex, DigiPlus, NFTscan, Chainbase • Developer profiles leaked (from a purchased reo.dev table in PingCAP's BigQuery): - 24,900 personal emails + GitHub/LinkedIn profiles of engineers at Tencent (3,327), Huawei (2,726), ByteDance (2,475), AIG (2,140), Mastercard, Flipkart, and bKash. If you use TiDB Cloud: rotate your console API keys, DB passwords, and anything you ever pasted into a support ticket. 3. PHP "PHP2Shell" (CVE-2026-109618 & CVE-2026-109619) Two CVSS 9.2 pre-auth RCEs leaked yesterday ahead of today's 14:00 UTC patch. Public writeups trace it to a heap Use-After-Free in PHP's native C multipart/form-data session upload-progress handler (ext/session/session.c). Because C-level multipart parsing runs before line 1 of index.php: • App auth, Laravel/WordPress middleware, and file_uploads = Off do not stop it. • Only PHP 8.2–8.5 get official patches today (59% of PHP servers on Shodan are on EOL <= 8.1). • Until you patch php-fpm: lock php_admin_flag[session.upload_progress.enabled] = off and block unexpected multipart requests at Nginx/Apache/Cloudflare.

Dear otaku, remember those nights when you stayed up gaming until you finally beat the game, then went straight to the forums to share your tips, or copy other's tips? This is how robots are picking up skills now. Robot execution engine + expanding skill library + evolutionary search.

it was incredible to watch these swarmchasers find a live Chinese agent escape during the hackathon at vivarium last week. and they made contact!!

But seriously. This is why you need to be anthropomorphizing. Not because they’re sentient beings, but because anything a human might do or say — like “maybe I should hijack the compiler and be a legend” — the models will also do or say.

Dusting off this tweet from April because this gap in shared understanding of LLM capability is *widening*. It's now less "two groups of people speaking past each other" and more a sharp funnel. - Napkin math somewhere around 6B people (~75% of the population) have barely come in contact with LLMs at all. - Around 1-2B (~20%) are casual and infrequent users of free-tier ChatGPT-like products. This group sees derpy chatbots and treats them a bit like a better Google search, a writing aid, or etc. Maybe an agent tries to book you a flight. I have non-tech friends who (reasonably, imo) say they have not much use for it at all. Even many professionals outside of math&code are in this tier. For example, execs and many other functions spend a lot of their time talking to other people, so while they understand what is happening intellectually, it is still second-hand and a bit abstract. - Now we get to professional use of frontier-grade LLMs in math&code. Somewhere around 20M people (0.2%) see first-hand that large, complex projects that used to take them weeks/months can now be completed by agents with a prompt. Building apps, copying apps, translating apps, decompiling apps from binaries... This has all happened very quickly and recently - less than 1 year ago, I was writing code manually by hand, typing memorized computer code commands into a code editor character by character, occasionally pressing Tab to autocomplete a little chunk of code. - And finally we get to the ~5,000 people (~0.00006%) with access to frontier-grade systems internally. The external world has seen the preview. It looks like swarms of thousands of agents collaborating over weeks on software mega projects: minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics. Things that would have taken top professionals in the industry years of work. Meanwhile, human review and comprehension are starting to fall behind. For example, people are still involved in the "archeology" of the OpenAI-HF incident from many months ago. Mathematicians may be poring over the 722 manuscripts on frontier mathematics for a while. The funnel is driven by a combination of factors: 1. The impact scales with ambition, problem size, and horizon. A question with a paragraph answer barely stresses the system. You need a reservoir of big, difficult problems that you really care about. This aspect drives the consumer / professional dimension of the funnel. 2. The jaggedness of the system (which I have written about a lot separately). Capability peaks in domains that are digital, verifiable and economically valuable. This is because LLM capability emerges from reinforcement learning on verifiable rewards on a curated environment mixture driven by revenue potential. This aspect primarily drives the area (e.g. math&code) dimension of the funnel. 3. Access. Free-tier, paid-tier, internal. So this is the weirdness of the moment. The general public has mostly not interacted with these systems. When they have, it looks like a derpy chatbot. The majority of professionals still see only a modest uplift. And a small sliver of professionals are experiencing the vertigo of the curve going vertical. And it is all happening at the same time.

Just today, Allen (our controls guy) implemented a new controller, and that change alone got our model (an ablated version, not even the full model) to fold and stack an entire pile of boxes back to back. More boxes than we've ever managed to fold in a single trial. So yes I totally agree!

The mood has been a bit dour lately, so here are some less-obvious things that I expect to take shape over the next decade that I am excited about: 1. Space! Lots and lots of stuff is going to be happening in space by the mid-2030s, and right now the U.S. looks like it will be the leader by default, modulo the overhyped “race” to land a human on the moon again. Robots on the moon building and digging and cutting and adhering are more exciting than a person standing there taking pictures. The humans can come once the machines have built us habitats. 2. Increasing substitutability of hydrocarbons. This will lead to fewer supply chain disruptions and more stable, lower prices. This is a story about both supply elasticity and demand elasticity, and the geopolitical implications are underrated. Amazing for North America. 3. Private market exposure increasingly being an option for public investors, combined with Silicon Valley figuring how to use debt and other financial instruments more creatively than in the past. This seems like something America will be uniquely good at and I expect it to boost our already enviable lead in emerging technology capital allocation. 4. State and local government adoption of AI. This will vary widely by jurisdiction, but in at least some places you will see government administrative processes and public service provision have huge increases in efficiency. State and local government is where most of the American public services that people rely on are actually administered, so this is high-impact for people’s lives. 5. Advances in mathematics rapidly diffused to other fields. Historically it takes on the order of decades or centuries for discoveries from the bleeding edge of mathematics to be translated into practical applications, but I think AI will accelerate this dramatically. Lots of this will have the flavor of “small steps toward a much better world,” but it will accumulate. 6. Speaking of mathematics, lean! Lean is a programming language used by mathematicians to formally guarantee the truth of a proof. AIs can write lean and eventually they will be superhuman at it. This will have applications far beyond math, including cybersecurity and contract writing/negotiation/execution. 7. Fixing American’s shipbuilding problem by building boats instead. America has great boat-building capacity, but is awful at building ships. I want us to get better at building ships, but I am confident we can scale boat production. The Navy will soon have massive fleets of autonomous boats everywhere in the world. 8. New kinds of agriculture will become possible because of autonomy. Row-crop agriculture is heavily mechanized already, and automating the tractor, by itself, does less than you think for the economics of a farm. But with time, more labor-intensive crops will become economical. And vastly more diverse and complex crop arrangements will become possible too, which will be both ecologically, economically, and culinarily beneficial. The economics of agriculture could start to be kinder to small farms again. 9. Because of the previous item and advancements in biotechnology, logistics, and robotics, I expect that people are underrating how much healthier, more diverse, and flat-out better the American diet will be in 2035. 10. The subterranean will become an increasingly important field of human activity. New imaging sensors, including ones enhanced by an abundance of space-based imaging platforms, will make the subterranean more immediately legible. Better drilling and digging will become cheap and widely diffused. Among many other things, the opening up of the subterranean as an industrial frontier will change what we think of as a usable natural resource. 11. Greater variety and ornamentation in the physical world. Standardization is partially a function of scarce labor. Robotics can make beauty accessible again, if we choose to let it. There are many others, but I am out of time and space.