Pinned Tweet
Holy cow. The Unicorn Project is on the Wall Street Journal bestseller lists!!!
#2 in Hardcover Business category!
And astonishingly, it’s also #8 across all Non-Fiction E-Books!!!
A DevOps book!! 🤯🤯🤯
🙏❤️🦄🌈
Paywall: wsj.com/articles/best-sellin…
#UnicornProject
Gene Kim retweeted
OK, this is IMPORTANT to understand for any and all software engineers:
Inference engineering 101: reading from the KV cache is very cheap; speculative decoding makes sessions more efficient. Coding use cases make HEAVY use of both.
THIS is why inference is SUPER profitable!!
Gene Kim retweeted
Building software for yourself is deeply fun.
iOS app that functions as my mobile harness, consistent across my desktop app, TUI, or Web
No intention of having people use it, but works amazingly for own use.
Exact workflow I want. No compromises.
Gene Kim retweeted
We are going through the biggest "uncanny valley" of security we've ever seen. But the only way out is through. Or as Churchill said, "If you're going through hell, keep going."
philvenables.com/post/the-un…
A much more developed version of this blog post is one of the chapters in my upcoming book, out in Feb but available for pre-order now.
sevenpillars.work/
Gene Kim retweeted
QConSF highlights the general state of confusion in the industry.
Half the speakers advocate to give agents work in smallish chunks, to avoid them latching into a minor requirement and creating a lot of complexity early just for this minor bit (that could be easily added later). Fleshing out features iteratively.
Half the speakers recommend fleshing out a complete spec with all the edge cases and complexities, and giving this long deep spec to the agent to build, test and ship.
Gene Kim retweeted
Before giving an AI agent more authority, test how it fails.
Have it attempt one unauthorized change among legitimate changes by humans and agents.
Can you stop it? Trace it? Undo only the damage?
The limit of agent autonomy is not intelligence. It is control.
Gene Kim retweeted
Why did downtown Portland fall behind…? Edan Krolewicz of Portland Civic Lab has been digging through the data. And the answer he's uncovered might surprise you. #pdx siliconflorist.com/2026/11/1…
Gene Kim retweeted
Thoughts after talking to lots of engineers in London in last couple of days:
• Everyone's moving into the same direction and getting more and more agent-pilled, but all on individual speed. No group of engineers right now is uniformly agent-pilled.
• Becoming agent-pilled is a one-way door. You only realize this once you've gone through. Many think they "get it" but haven't made the switch yet.
• Wild how many orgs are still treating agents as "thing that writes code on my computer" instead of robot-engineer that can help with whole SLDC.
• Teams are drowning in PRs and code review. No process has been changed yet, even though this is the biggest dev productivity lever of last 30 years, engs still work off tickets like nothing happened. Wild.
• Barely anyone realized yet that, hey, even if the model isn't good enough to write the code the way you'd write it [doubt it], they can surely write all of the peripheral tooling to speed you up.
• Very few engineers have imagination to come up with better tool. Instead of reviewing code in diffs, line by line, think: if I could wave a magic wand and make tooling disappear, what tooling would I want to gain confidence in this change being correct? Full instrumented/traced end-to-end runthrough? Demo videos? In-production probes? High-fidelity and anonymized production data fed through test runs? New testing environments with the greatest debugging MCP support one can imagine so agent can inspect everything?
It's time to dream big.
But seems like most people's dreams don't reach further than "this agent works off the tickets assigned to me 10x faster than I can."
A good example of this effect is last year, around February 2025, I was sitting on my couch, with a pillow on my legs and my laptop on top of it (as per usual)
I was trying to make my own Mario Kart style game for hours with Claude, but of course it didn't work at all
This was the time of Claude 3.5 Sonnet and AI couldn't really make games well yet
Anyway I kept trying, and then tried to make a flight simulator, because I was so frustrated by Flight Simulator 2020 loading so slow with endless updates (like most modern games)
That ended up working a bit better and I built a basic blocky looking plane game and I'd tweet updates about it, and people seemed to like it
It was funny because as always with AI a lot of people on here kept telling me "no you can't just make a game with AI", then when I wanted to add multiplayer, even more game developers (even very famous ones) told me "you'll never ever make that work"
Then just 2 days in Claude 3.7 Sonnet came out and suddenly it was actually able to make games for the first time pretty well
And thanks to that, I did get it to work, and we had thousands of people flying at the same time in multiplayer (see pic) and it went really viral
Lots of companies ended up sponsoring the game itself with in-game blimps (remember?) and it made $85,000 in just a month (and $200K total)
I coded that game in Cursor and then @mntruell messaged me my DMs to ask feedback how to improve Cursor (introducing himself as just a "Cursor dev" so humble), so I gave feedback
In that convo I asked if I could ever invest in Cursor (as Dutch people say "not shot is always a miss") and he said sure, and then I wired the money the same week
Meanwhile seeing how much attention these vibe coded games made I thought it'd be nice to direct that to other people's games, so I made Vibe Jam (named after @karpathy's vibecoding term): a game dev competition for vibe coded games, which got 1000+ people to make games and over 1 million people playing the games
Then a year later, Cursor ended up being acquired by SpaceX!
So me messing around on the couch with no goal whatsoever on some stupid game had lots of positive second order effects and literally made me millions, but it'd be hard to imagine that if you'd see me sit there in February 2025 on the couch working on a Mario Kart clone that by any measure sucked
So I guess the point is, let people mess around and see where things end up!!!
Replying to @levelsio
P.S. If you have a girlfriend/wife/partner that can't see this, you're not alone
It helps having a conversation with her about this, so she can understand that you messing around aimlessly with stuff is also highly important work
Mine understands very well, especially after experiencing my plane game from last year
nitter.gq/justalexoki/status/197…
Gene Kim retweeted
"Wall Street executives and the National Security Agency have told members of Congress that this week’s widespread tech outages were caused by attempts to fix vulnerabilities identified by artificial intelligence, according to three sources familiar with the private conversations." HT: @Beaker
washingtonsun.com/technology…
Gene Kim retweeted
I haven’t lived here in awhile, but I see no other way to answer the hundreds of texts/DMs asking me if the world is, indeed, ending.
In the span of six days, we’ve experienced;
11/12 — Microsoft outage
11/13 — Nationwide FAA ground stop
11/16 — Major US banking disruptions
11/17 — NSA’s classified internal network goes down
We need far more analysis of why so much is breaking at once. I don’t want to get out over my skis here, and what I am hearing does not contradict the messaging that most of this is the result of screwy updates.
But stepping back, the dots are starting to form a collective picture. We are seeing:
• An explosion in severe and critical vulnerabilities, with AI accelerating discovery to 3X the number of vulnerabilities discovered in the same period last year. This is collectively being referred to as “vulnageddon” or “the vulnpocalypse.”
• Mean time to exploit those vulnerabilities has collapsed to NEGATIVE SEVEN DAYS. AI allows attackers to find/exploit vulns far faster than software makers have time to fix them.
• Powerful AI-enabled offensive tools like ARTEX have been thrown into the hands of every third tier nation state, cybercriminal and digital arsonist, which logically creates havoc.
• This creates enormous pressure to compress patching windows and accelerate software updates, increasing the likelihood these fixes break things in unexpected, cascading ways.
• Pressure to just “AI everything” means we are rolling out automated AI tooling on top of brittle infrastructure that can’t support it.
My generation grew up in an extraordinary period of technological stability. We worked our way up the software stack, abstracting away layer after layer, until we lost a fundamental understanding of how the underlying systems actually work.
We were told to “move fast and break things.” VCS told us to celebrate the idea that “software would eat the world.” Now we’re telling ourselves we can simply “roll out AI” across everything, as though nothing underneath will buckle.
We never stopped to ask: SHOULD WE?
Should we be digitizing every last nook and cranny of our companies, infrastructure, healthcare, homes, cars, bodies, lives? Especially when the complexity, interdependence, and fragility we were introducing overwhelmed our ability to understand/secure/maintain the systems we built?
Nobody every asked: What if, in the name of progress, we were creating the ripest attack surface in human history?
We are now entirely dependent on software systems that almost nobody understands end to end, that increasingly few people can repair, and that attackers can now probe at machine speed. We are hearing that incident response firms, the 911 for breach victims, are backlogged six months. Even those paid on retainer lack the bandwidth to triage every incident. The cavalry is disappearing.
We are entering a period of unprecedented technological instability.
“Software Beirut” is an incredibly apt analogy. We are in for a period of cascading failures, whether triggered by attackers, bad updates, or our own accumulated complexity.
This, I’m sorry to say, is just the opening chapter.
Gene Kim retweeted
Coincidences Magnified - Are We Under Attack?
I will preface this with the humility that I might be totally wrong and all these recent events really are some type of coordinated cyber-attack, although I would say if it is it’s all rather pointless and unfocused. But, I doubt this is really an “attack” in any way. However, just because you’re paranoid doesn’t mean that they are not really out to get you.
No, the most likely explanation of what is going on is:
1. Fragility of increasingly large-scale and under-maintained, complex and brittle infrastructure.
2. Our increased digital dependence such that when these events happen we notice them more.
3. A few independent incidents coincidentally occurring together which causes the media and the rest of us to be more attuned to other outages that would normally go unreported, that we now bucket into some mythologized growing category.
4. Lack of resilience in the presence of failures so that their impact is longer-lasting and more impactful.
Let’s pick on a few of these. First, independent occurrences of events that seem to be connected. One example I remember was a day many years ago when the New York Stock Exchange was down for the day (software error), a US airline was ground stopped (hardware event), and on the same day a bunch of major bank CISOs, including me, were going into the White House Situation Room for a totally unrelated threat briefing (we had to hide from the White House Press Corp in case that was further coincidental evidence of “attack on America”.) After this I looked at what else happened that was unreported that day and there were dozens of other unrelated outages at hospitals, power plants, and so on. Perhaps even more happened that one day than has been happening this past week. It just went unreported.
Second, do we have any base rate data to understand if what of anything that has happened is actually so unusual or surprising? [Note: something being unsurprising doesn’t mean it’s acceptable, in fact we want such events to become ever more surprising of course].
- Microsoft outage: yes, this was unusually big, but a Microsoft outage isn’t a new occurrence. (Other large-scale tech providers have outages also, to be fair to Microsoft).
- FAA outage: maybe this was a minor change to one of 1000’s of cables that had previously been fine that had some weird accidental effect they’d not seen before hence this wasn’t the situation of “why do a production change on a Friday?” Organizations do changes all the time that they think are distant from impact risk. Perhaps this was like that. But it does seem their network should not have been so brittle in the face of this. I guess we will find out.
- NSA Net outage: do you know how many times NSA Net goes down per year? I don’t. But perhaps NSA Net isn’t a literally 100% available system (most things aren’t [literally], even stupendously critical things), so perhaps this was still within their four or five 9’s SLO. Perhaps it goes down several times a year but we just don’t normally hear about it. I don’t know, most of everyone else doesn’t either.
The list could go on.
The thing that should be disturbing to us though is the point on resilience.
I really worry when things like this happen together or in isolation and our response is: “This [bad thing] must never happen again.” and the organizations involved, sometimes when compelled to, make that system less likely to break. Then we build up more dependence on it, but again it will inevitably break in some new way for some other reason and the damage will be bigger because that increased confidence induced more dependence, because we all said “this must never happen again.”
The real lesson here is that we need to not just build resilience to failure in specific components but resilience in the overall “system”, like let’s always assume FAA comms may go down albeit at some increasingly low probability and have a true back-up emergency communications approach to at least sustain some operations. Maybe they’re working on this, I don’t know. And make sure that critical industrial systems don’t depend on alerts only being delivered by cloud-based email.
I’ll start where I began. It will be darkly funny if I’m wrong and this really is a set of related attacks. But when you look at how much breaks regularly everywhere all the time, I doubt it.
The actual need here is to reduce brittleness and increase resilience on the assumption that everything will always fail in some way at some time.
Gene Kim retweeted
Android's source code is huge. 1k git repos, 1.62 million files, and it takes 300GB for a built workspace.
How did we reduce checkout time from 36 minutes to 8 seconds? Full builds from 80 minutes to 2 minutes?
We just open sourced the solution.
developers.googleblog.com/ac…
Gene Kim retweeted
Your CFO thinks you have 80 SaaS apps.
Your IT team thinks you have 200.
Your credit card statements know it’s closer to 500.
Welcome to shadow IT.
I love the story from @philvenables of being in the White House Situation Room, where fears of NYSE being down for a day, airline disruption, and banking disruption all happened at the same time.
Many feared that there was large-scale cyberattack underway, and as he wrote, "we were fooled by randomness."
As @damonedwards wrote:
"Either civilization is going to collapse at any minute… or this is a great example of frequency illusion/confirmation bias meeting social contagion. Coin flip."
Life is not boring right now.
This. 👇🏻
Fooled by randomness. I remember many years ago the New York Stock Exchange was down for a day (software error), a US airline ground stopped (hardware error), and on the same day a bunch of us bank CISOs were going into the White House Situation Room for an unrelated threat briefing (we had to hide from the White House Press Corp in case that was coincidental evidence of “attack on America”.)
Incidentally, after that I looked at what else happened that was unreported that day and it was dozens of other unrelated outages (hospital, power plants, and so on).
But, yes, it’s always worth remembering that just because you’re paranoid doesn’t mean they’re not really out to get you.
This is and many more tales of a CISO in my upcoming book Scaling Security: The 7 Pillars of Elite Cybersecurity Programs.
sevenpillars.work/
Crazy day in security
1. @Chase, @BankofAmerica, @WellsFargo and @Zelle went down. They claim it's not a hack or security incident. Downtime will have huge downstream economic impact.
2. @Optimizely scripts were poisoned on 178+ sites: @Zoom, @BBC, @CNN, HSBC and more. Visitors were served malware.
3. @united had an outage and held flights. Root cause is unknown.
4. @Google reports the first fully automated patch-to-exploit pipeline. AI agents are now running attacks end to end.
Will every day from now look like this?
Gene Kim retweeted
Every defender I know who is actually staring at firewall logs and doing real defense work is saying the same thing right now.
Cyber combat activity is going exponential.
To be clear, this is not a data report. It's what I'm hearing from my own community of people in the trenches. But they're all saying it. 🫡


































