(Honestly, it spends more time in uefi preboot than it does to actually boot the system.)
The problem with vibe coding isn't the AI, it's how people use it. It's that people with no idea of what they're doing, now think they can do what used to require an expert.
Edit: I've just tried to log in and it's failed through both Google and email. Whatever they've done, they haven't tested it well.
It's up to the people involved in the process, to gatekeep AI to not *ship* bad software.
I have a bunch of plants on my window sill, and considered getting moisture sensors and automating watering, but I realised that kind of defeats the point of looking after the plants myself. On top of the satisfaction of manual work, I recently realised that it's taught me an intuition - the quirks of how much water each plant wants at different times of year (even cacti are thirsty in summer!), and how soil composition and pot size affects water retention. It's probably not generally transferable knowledge, but it was fun to realise I'd come to know my plants really well.
Maybe your opinion is that I'm not vibe coding hard enough, and if I did AI better we wouldn't need code reviews. I disagree - both engineers and LLMs are still adding unneeded complexity or choose the wrong abstractions in ways that will cause problems down the road, don't think about all the corner cases, and pick the wrong handling for error conditions. They add in things that seem like a good idea to them and don't stick to the spec.
If my company stops doing code reviews for the software that matters, I'll quit.
Engineers not in the startup crunch should aim for using LLMs to code better, not faster. Faster just creates technical debt and ups your long term costs.
OK, now I'm sure that the guy is full of shit.
IMO they’re primarily about ensuring shared understanding and agreement, not just about how things are changing now, but how they’ll change in the future..
That ship has sailed.
Huang has a problem now because his ability to advocate for NVIDIA in a politically charged environment is damaged.
He can't be an effective CEO for NVIDIA if there's a plausible (even if unproven) suspicion that he's on a geopolitical adversary's side.
He was very strongly advocating for selling to China before this. At that point, it could be understood simply as him looking out for NVIDIA's business interests. Now, there's doubt sowed in everyone's mind about whether it really was just that.
Maybe you don't care, because you think selling to China is the right thing to do - fine, there can be different views on this.
But, a lot of influential people very much do not hold that view, and he needs to be able to effectively communicate to them also.
He's operating in a highly charged political environment in which his credibility is a key asset. He can't effectively advocate for NVIDIA's interests if his credibility is damaged.
And this goes directly to his credibility on the most politically charged issue of all: The geopolitical rivalry between the US and China, and NVIDIA's role in it.
I agree it's a tragedy for NVIDIA, he seems to have been exceptional as a CEO, it's just too late now.
NVIDIA is licensed by the United States government to sell GPUs to China, and they do so at a rate well under their licensed quantity.
Read up on 91 CFR 1684 which directly contradicts your claims.
Your most important argument is categorically false and completely unresearched.
Your subsequent claims also have no basis: you cite a bunch of cases in which people leaked sensitive information as equivalent to the Jensen case, where the affair was with the CEO of a customer they were legally selling product to, and where we have no evidence that any leaks occurred.
The US Government is using NVIDIA chips for all sort of stuff across all classification levels. Subsequently, Jensen has deep relationships within the government that warrants additional scrutiny.
Him authorizing his company to sell massive amounts of chips to a restricted country is massively illegal, but companies work around that and have since forever. Doesn't make it less illegal, but that's only a small part of the problem.
The real problem is that nobody knows how much sensitive info Jensen gave this woman or how far that info spread.
It's really about fairness. Us technical salespeople take training that tells us all of the ways that deals with restricted countries happen and how we will immediately get fired ten ways to Sunday if we forego that to close quota. Also, normies with clearances tell their partners stuff they shouldn't. Stuff they think isn't really sensitive since it seems so minute. This isn't bad until said partner spills at the wrong place and wrong time. That's how normies get sent to prison for treason. Loose lips sink ships and all that. [^0]
Getting fired -- WHICH HE ABSOLUTELY SHOULD BE -- is the least of his problems.
[^0] Examples:
- https://www.ctvnews.ca/world/article/retired-us-army-officer.... Army officer leaks secrets through dating website to sound cool; gets six years
- https://www.foxnews.com/us/pentagon-leaker-jack-teixeira-rec.... Leaked secrets about Ukraine war through Discord to sound cool; gets 15 years
- https://www.westernjournal.com/ex-cia-officer-accused-leakin.... CIA officer leaks stuff to journalist girlfriend; gets 10 years
This is exactly how it played out, too, which is why Huawei chips have started taking inference workloads in China and we're now starting to see them in training too.
But I wholeheartedly disagree with ousting Jensen. And given the timing of all these releases, I wouldn't be surprised someone inside the government wants to bring him down.
I think there are a lot of people who want to bring him down. His competition - AMD, Broadcom, Google. OpenAI and Anthropic also have advocated for banning chip exports to China because of Chinese lab LLM competition. Then there are both Republican and Democrats who likely support the views of OpenAI/Anthropic.Finally, because he's of Han Chinese origin, I bet they thought they could get the media to coordinate a narrative around him that will force him out. Of course, he sort of made it easy to pin something on him.
There are a lot of Game of Thrones going on in AI. Everyone is jockeying. Everyone has different incentives. Super high stakes game.
I am not entirely sure I buy the Spy thesis. And Alice's husband knew this all along. Which makes the whole thing more doggy.
And Jensen's view isn't wrong. There is no way you could contain China via import restrictions. Not when they are the second-largest economy and still rapidly advancing. There are better ways to handle this, one huge ban isn't one.
But I wholeheartedly disagree with ousting Jensen. And given the timing of all these releases, I wouldn't be surprised someone inside the government wants to bring him down.
And anyway, whether Nvidia created CUDA and deep learning are all moot at this point.
This is factually incorrect. It is entirely legal for companies in Europe to restrict relationships where one party is a direct subordinate to the other. They can also legally require disclosure of such relationships.
This is common because these relationships can create significant conflicts of interest.
As someone else in this thread noted, the former CEO of Nestle (a Swiss company) was fired because he had a relationship with a direct report that he failed to disclose, and the former CEO of BP (a UK company) resigned because of a similar situation.
Elon regularly gets away with way more morally (and perhaps legally) egregious stuff and yet he has increasingly cemented his position. Partly through indirect leverage from his companies and partly from buying access to power directly.
Even Sam managed to stage an employee revolt to stay in control.
For founders of companies perhaps the lesson keeps repeating itself - keep control of your companies no matter what.
I’m not sure this is good for shareholders and the larger AI ecosystem either. Corporate governance is obviously important especially for mature industries.
But Jensen has consistently demonstrated ability to keep Nvidia alive during dire times and eventually delivered the biggest profits in corporate history.
Edit: some cynical explanations:
1. China hawks in tech and admin wanted him out
2. Nvidia competitors found oppo research too juicy to ignore and sequenced a set of destructive reporting
3. People in AI found Jensen’s grip of the supply chain too annoying and wanted to loosen it somehow
4. Internal politics / someone wanting Jensen’s job
5. Some combination of the above
While I appreciate the spirit of this comment, I think that this ship sailed long time ago when United States elected a convicted felon twice, who then proceeded to virtually break all the laws with no repercussion.
In Europe, workplace relationships are considered a private matter as long as workplace peace doesn't suffer, and a ban would be considered legally impermissible. Europeans manage it with disclosure requirements and reassignment, not bans.
He had an affair with a woman believed to be a Chinese spy, who it appears smuggled potentially billions of dollars worth of Nvidia chips into China. And now he is alleged to have slept with an Nvidia executive who reports directly to him.
If he had slept with a waitress or his friend's wife, he would survive this. Instead, he chose to have exactly the type of indiscretions that get you fired.
Most companies have policies that strongly discourage this, or forbid it, and for good reason.
It's a big enough deal that former CEOs of Nestle, Norfolk Southern, McDonald's and Intel were either terminated or resigned after similar relationships were uncovered.
The scenario in which dual-share structures can be a problem is if things don't go well. Specifically they greatly limit the shareholder's ability to protect their investment by removing the CEO. That's why it's considered poor governance.
1) The fact that he cheated on his wife isn't particularly relevant, other than the lack of a prenup meaning she owns half his stock (4% of the company).
2) The fact that he slept with a subordinate shows bad judgment, but isn't a huge deal (although improper and opens up potential quid-pro-quo lawsuits)
3) The fact that he slept with a Chinese spy is a huge deal.
4) The fact that there are allegations that he was aware that they were selling illegally to China is a huge deal.
5) Even though those export control were arguably bad, the law must be respected and the being seen ignoring the law would have a chilling effect on appearance of a fair democracy.
6) There is definitely a demonstration of lack of judgment (getting caught selling to china rather than maintaining plausible deniability), (sleeping with a spy rather than just hiring an escort)
7) Despite all of this, in theory he would arguably still be the most profitable CEO (if he doesn't get charged with a crime, AND if there are no other issues to discover) in the medium term, but where there's smoke there may be fire. Does this raise the odds of circular financing and other shenanigans?
Honestly this arrangement (total comp aside) is how it should be for all these companies. When Facebook went public the dual class structure (and Zuck retaining voting control) was a big point of contention.
Also having been more capital intensive and founded further back, Jensen had less sway in the early days of the company and may have put off using the growth to his advantage.
> For Ms. Huang, there was reason to celebrate. Backed by wealthy investors in Shanghai, she had just restructured a company that would soon buy billions of dollars in Nvidia artificial intelligence chips. The U.S. government had barred Chinese companies from buying the advanced chips, but Ms. Huang had sidestepped this prohibition by creating a firm, Megaspeed, through a Singaporean shell company.
> Her close relationship with Nvidia’s leader also appeared to smooth the way. Four days after the ownership change Megaspeed was named a Nvidia cloud partner, giving it access to the coveted components.
If NVIDIA doesn’t fire him, I have a feeling a lot more women from a lot of other companies may be showing up with various proposals.
In fact sleeping with a subordinate (look up Kimberly Powell) creates substantial legal risk for the company on its own, even without all the other stuff.
> We don't even know what arrangements him and his wife may have.
He apologized to his wife and said that his actions hurt her and that there was no excuse for what he did, soooooo, yeah we kinda do.
Poor judgement is the kind of shit Zuckerberg or Musk do, and their poor judgement actually impacts the whole world in big ways, but they can't be fired because they don't give their board any control.
Jensen's management of NVIDIA has been good, and NVIDIA has generally been a great force for real innovation without much rent-seeking or corruption like the others. They haven't started messing with politics or trying to buy elections. It's kind of sad to see Jensen be the one we topple at a time of rage against corporate power when he's probably one of the better stewards.
Maybe his idea of asserting control is less about voting rights and more about family allegiance.
As bad as cheating is morally, we don't really know how his wife feels or if they had some open relationship type arrangement, and it's not really material to his running the company because it's been doing very well during all of that. I don't think it makes much sense to fire someone over this, it's more like a power struggle with other parties vying for control.
Without Jensen Huang's vision, Nvidia might have remained primarily a graphics chip maker, and the deep learning revolution would not have arrived when it did. Long before deep neural networks took off, Huang backed the use of Nvidia GPUs for scientific and general-purpose computing, including the development of CUDA. In 2012, Krizhevsky, Sutskever, and Hinton used Nvidia GPUs and CUDA to train AlexNet, whose breakthrough performance helped ignite the modern deep learning boom. Nvidia's investment in programmable GPUs helped make that breakthrough possible.
The critical vision was that GPUs could become a platform for computing beyond graphics—and that Nvidia should invest in making them programmable before a large market for that use existed.
The board should punish Huang without punishing the industry and the nation. For example by cutting his salary and stock to 0 for 2 years.
- His board setup is ... odd. He isn't Chairman of the Board (a role over 50% of Fortune 100 CEOs apparently hold). He seems to be responsible for Nvidia having a wierdly more egalitarian structure (all directors including him, are equal, with a designated lead independent director)
- His compensation (~$35M, primarily stock) is ... not up there with other tech CEOs
- He doesn't seem to have ever pushed for super-voting rights (he only has like ~3.5% of the company), but I guess that's hard to change once you IPO
I kinda wonder if this ends with him taking a month off, spending 12-18 months transitioning to a successor and ending as the chairman of the board or some such. He really does seem vital to the company (and thus the stock price).
At the time I didn't think it would happen because of impact on Nvidia's stock price, but at that point we didn't know about his second affair - with Kimberly Powell.
I'd suggest this latest revelation makes his position untenable, especially if the board found out about the second affair via a source other than Huang himself. He should have spilled the beans on that when the first affair came out, not that it would necessarily have saved him, but withholding the information - if that's what he did - should condemn him. It's another serious lapse in judgement and a failure of integrity that is not befitting of one of the most powerful corporate leaders on earth.
Goodness knows how his wife is feeling at this point. Certainly makes his apology from a couple of days back ring rather hollow.
EDIT: The fact that someone like Jensen Huang - who I'd previously considered to be an incredibly smart guy as opposed to a nitwit - can somehow completely implode his life in almost the most ridiculous way imaginable does give me some hope that at least some of the troublemaker ne'er do wells that have wielded so much power over us all over the past decade or so might also experience some just desserts in the not too distant future.
I think the real story is not the affair, which was already reported in asian media in 2024, but what specific catalyst broke that protective shield and made continued containment untenable.
He's had the job since before GTA 1 started development.
I don't think it's good for Nvidia if everyone has to go through a few labs APIs to get access to AI inference.
It makes sense to me why they would benefit from chinese labs having more compute.
Why not just sell to not-sanctioned entities, ensure compliance to the best of your abilities and otherwise keep your head down? Again, wasn't like they struggled to find buyers for their chips.
Considering Megaspeed, I now see this in a very different light. Him cheating is personal and I wish the affected members of any family of a cheater the best, but not relevant for us outsiders.
Being honey trapped very bad securities wise, but he'd be a victim (and evidence for why we shouldn't discount Mistral in the frontier lab race [1]). If sanctions avoidance went all the way to the top, that is hard to negate even for a board that is interested in keeping him. At that stage he was neither honey trapped nor "merely horny"...
[0] https://www.youtube.com/watch?v=1H3xQaf7BFI
[1] https://time.com/archive/7146565/hollandes-popularity-goes-u...
Dual-class share structures have become the defacto standard for founder led companies, giving founders disproportionate (even majority) voting power.
Someone like Zuckerberg effectively controls the board and would be extraordinarily difficult to oust against his wishes (absent some exceptional legal intervention or a decision to step aside)
Affair of a CEO with a direct report and member of senior leadership? Kind of a big deal
Affair with a Chinese intelligence agent? Quite a big deal
The no-going-back issue is the corruption. Jensen corruptly quashed an investigation into Megaspeed. He then quite clearly expected Republicans in Congress to back him, which is why he only came clean when forced to.
One of the greatest failures of Nvidia's Board was in failing to push for a successor. (Musks' companies face a similar risk.)
Why is this? If you go look at the financial statements of airlines (for which Airlines for America is the industry trade group), you will understand that a large portion of their profits come from monetizing human friction and attention.
Imagine things like upsells for their branded credit cards, feeding into the points game, etc. These are all things that appeal to humans, but that agents are almost certainly going to skip and then relentlessly price-compare the flight at the end of the funnel.
Now, once Amazon’s appeal was denied by the court, it became obvious that they needed to resort to technical solutions if they could not block it by trying to leverage the CFAA. Therefore, these industries are now implementing technical solutions.
Ultimately, personal agents are a huge consumer benefit, but they challenge the business models underlying much of the internet. I expect to see more of this kind of behavior from incumbents, which ironically leaves them vulnerable to disruption by players who do support agents, as this market shift plays out.
It is possible to detect Claude controlling your Chrome browser, as well as many computer use agents. Give this a try: https://demo.switchfrog.com. Ultimately though the mainstream agents will (and mostly already do) self-identify and businesses will adapt.
Source: I am the founder of a company in the space, Switchfrog.
This is a small world. The US has something like ten major airlines. Eight if you don't count subsidiaries. Someone at the first one made a decision to embrace an exploitative pricing model. And then, someone at the second airline looked at this and thought to themselves, "you know, we're not gonna organize a PR campaign to stigmatize it, we're gonna try the same".
[1] https://mashable.com/article/airlines-set-prices-artificial-...
Thats a DoT requirement all flights to, from or within the US. (7 or more days before departure)
Cancelling a United flight is one of the easiest cancellations I've dealt with, their app is one of the best but its just as simple from the website
The airlines use every nasty practice they can to raise costs to you. Hidden fees, raising costs by just viewing flights and all kinds of other crap.
Let them burn.
I often use Kayak or google flights to screen multiple airlines, but once I've found a flight I like, I generally book it on the airline's site, and the UI is basically identical to Kayak. Put in some dates, some locations, click search, browse through the options, pick the one you want. If you are flexible on dates, there is a little +/- 3 days checkbox that will show a table of flight prices.
I'm probably stupid, but what are the improvements that would help customers find the best deals?
Booking a flight is obfuscated and you'll have to comparison shop across multiple websites or even devices to get the better deals.
Same with hotels - they have tons of rate obfuscation and you have to cross shop their individual websites and the aggregators as well as third party discounts to figure out best deals.
Concerts are similar but different. Sometimes they should be selling tickets via lottery but make everyone fight over them. Other times there are presales. Other times you have to shop through scalpers. Arguably tickets are less painful than the other two, but they're still a manipulative mess where you waste a lot of time.
They should meet the customer where they are instead of being antagonistic to their customer by default
- already streamlined (spend 2min on Google flights, seatgeek, etc...); we aren't stuck with early web UX and latency
- I want to have 100% control of the outcome or would have to write a long list of preferences and rules (middle seats, layovers, etc...) to feel good about trusting automation
- require making a large payment. the agent is not getting free reign to use my cc
- give me some personal satisfaction and enjoyment - its delightful to look foward to a trip, concert, etc... and planning is part of the fun and anticipation imo
Sure, the airline may not notice when it happens but it can cut those processes short when they have a full ban.
If this is indeed the case -- good riddance to that practice! Why should I be sorry that it goes?
At this point it feels like I have an adversarial relationship with many merchants I use online. I wish the simple “you have a product, I have money” relationship was possible. I don’t want to take part in your giant human behavior optimization problem. You have a product and I have money why does it have to be more complicated than that?
Realistically though I know that’s not happening. I will continue to access online merchants in the way I prefer and will do my best to avoid detection or blocking. At least so far Claude is happy to play along but I suspect I will need to seek out an obliterated agent in the not to distant future. (Once the corporate capture of the frontier cloud AI models is complete.)
I guess the best option is, to use a good reseller/travel agency that provides an API/MCP for booking and managing flights.
For finding the cheapest deals agents can still be used. Just the booking itself is probably better done manually. It also eliminates the risk that the agent does something stupid.
edit: they might try to block agents, but this won't be possible with browser use. they can always throttle users by IP or fingerprint. Letting an agent book a flight might be risky, because they might cancel a booking or impose a fee if they somehow detect it.
Just to be clear, your argument is that it's "counterproductive" and a "local optima" for people to have complete and accurate price information available for purchasing their tickets? And that making this information harder for them to find is better for them? Can you point to any real-world examples of an asymmetry in pricing information ending up saving the side with less information money rather than the side with more information to use that in a way that disadvantages the other side?
I was pretty frustrated recently when I got a flight for 20% less by clicking through from a Google Flights link than I did from searching directly on aircanada.com while logged in with my Aeroplan account. I know exactly why it's the way it is, but it's pretty discouraging to feel that loyal-customer behaviour like going directly to the airline instead of an aggregator is punished with a higher price.
What if the airlines instead focused on delivering great service at a competitive price?
"Nah, let's escalate this arms race instead."
The joke is that people wanting to write Python like Java - with an implicit reference to an object's instance - would attempt to import Java's `this` keyword.
Instead of getting Java-like behavior, they're presented with the Zen of Python. You wouldn't be wrong to think that it was as much a commentary on other popular languages of the day as it was a guiding ethos of Python itself.
"Beautiful is better than ugly. Explicit is better than implicit."
The reason that's especially interesting is that I was using an AI to write something that needed that functionality, and it spewed out that same Python code, which I recognized immediately. It was slightly modified with respect to naming and comments because, I assume, it was stored slightly imperfectly in the model's weights.
I like reading his answers. He always seemed to have something interesting to say, and his answers are kind, light-hearted, approachable and insightful.
(https://en.wikipedia.org/wiki/Zen_of_Python has more about the discrepancy.)
He was always so willing to foster discussion that nearly every interaction resulted in an improved understanding of some topic and a shared eagerness to try or learn more out something tangential that came up via digression. I never knew how many types of voting systems there were nor how extensively they are researched! He also reached out about a comment [0] I made a few years ago which made me feel less discouraged and more welcome.
Tim's writing style was greatly appreciated by myself and many others. What he had to say was often lengthy, to address reader assumptions or some nuance, but was written with such a jovial tone that it was rarely dull or inaccessible.
RIP Tim
And after that three months suspension BS drama Tim just got back to work on Python. Saintly.
btw. made me stop having a favourite programming language.
You never know when an email is going to be the Last Email. Don't put off reconnecting with old friends.
(EDIT: Robert Woodhead apparently shares in the credit, too: https://www.ceas.cc/papers-2004/136.pdf https://ceas.cc/papers-2004.html . Apparently Rob Hooft was another implementor? I can't find a good source for that though.)
For me, Tim Peters was that guy. I first met Tim in person at the first PyCon at GWU where I rode an elevator with him and Barry Warsaw. Let me tell you, it was a wild ride. He was a genuinely really nice, smart, kind guy.
RIP my friend.
Another thing that made me feel close to him was his work on the Python Cookbook 2nd Edition, in which he (and other big names in the community) delightfully talks the reader through intricacies and insights of programming, "We read Knuth so you don't have to".
Watching Tim do his thing was a pleasure. He'd go from long, deep technical dives (that he'd make as approachable as possible) into comic relief and human connection in the fractional wink of an eye.
The brief interactions we had will always be cherished as high points in my Pythonista trajectory.
One favorite of mine was this response: https://www.python.org/doc/humor/#id15. I also like https://www.python.org/doc/humor/#id19 very much.
I was an active participant in the comp.lang.python Usenet channel in the late 1990s.
I learned much from reading Tim Peters' posts. To many, including me, he was affectionately referred to as "Uncle Timmy."
One of the things I learned was that I still had much to learn about the art of programming.
Much of the little I know about numerical methods comes from reading his posts. It was fun to read his stories about Kendall Square Research and the (for me) old days of computing.
I remember posting questions about things I didn't understand, which he figured out and explained. [1]
I remember bringing up concepts I thought were specialized and esoteric, only to find that it was actually just my lack of experience. He gently corrected misapprehensions like mine. [2]
He wrote a lot on comp.lang.python. If I recall correctly, at the 7th International Python Conference in Houston, November 1998, there was Festschrift in his honor. He would use <wink> in his messages, and often end with something like "let's-hear-it-for-lambda-in-curly-assignment-stmts-ly y'rs - tim" (that's from the posting where he outlines the Zen of Python). I wondered if they were collected by looking for messages containing an '-ly' only to have Andrew Kuchling remind me about the "From" header. <wink>
Here is Guido van Rossum starting the thread on "In Memoriam: Tim Peters", at https://discuss.python.org/t/in-memoriam-tim-peters-1952-202... .
Here's an interview of him at PyCon 2006, https://www.youtube.com/watch?v=1wAOy88WxmY
[1] In one case, I was parsing a file format where a field was usually three spaces, but was sometimes an integer. I couldn't figure out why one parser was so much faster than the other. The answer was that one tested for three spaces first, and if that failed tested for the int, while the other did the int test first.
Yes, this is something a senior developer knows. I bring it up now to highlight how hotshot-me didn't have that experience, learned that experience, and learned it from someone who was kind about sharing his knowledge.
[2] I was doing bioinformatics back then, and learned about dynamic programming as way to do sequence alignment. I didn't think others knew about the technique, and thought I was bringing esoteric new information to the discussion. Uncle Timmy had used dynamic programming techniques for speech-to-text software at Dragon Systems, and politely pointed out it was more widely known. He also implemented Python's 'difflib'.
$ python -c "import this"
The Zen of Python, by Tim Peters
Beautiful is better than ugly.
Explicit is better than implicit.
Simple is better than complex.
Complex is better than complicated.
Flat is better than nested.
Sparse is better than dense.
Readability counts.
Special cases aren't special enough to break the rules.
Although practicality beats purity.
Errors should never pass silently.
Unless explicitly silenced.
In the face of ambiguity, refuse the temptation to guess.
There should be one-- and preferably only one --obvious way to do it.
Although that way may not be obvious at first unless you're Dutch.
Now is better than never.
Although never is often better than *right* now.
If the implementation is hard to explain, it's a bad idea.
If the implementation is easy to explain, it may be a good idea.
Namespaces are one honking great idea -- let's do more of those!It's perfectly sufficient to have a CVE for just memory corruption, so the "RCE" framing here feels overstated. Particularly when an LLM should easily prove RCE out if it is in fact present.
>As of the time of this blog’s publication, no exploit is available and there are no reports of exploitation in the wild.
Since when does "memory corruption" = "RCE"? Last week I found and reproduced a memory corruption (a use-after-free) in PHP's runtime after running an on-premise agent for just 30 minutes. There's nothing about it on the internet. Should I write a blog post about a scary new vulnerability? :)
On the other hand, many people claim that machine-generated software STINKS. That it is extremely verbose, that it reimplements similar methods over and over again. That it would rather cut corners at a low level than build high-level abstractions. That it pulls in libraries and makes serious security mistakes. That it writes tests that are basically incomplete. That it is unable to synthesize and keep context. In short, it has the same flaws as humans.
I think these are binary points of view, and both are right. We have vampires that are incredibly fast, super-strong, super-intelligent and super-agile, that suck your blood (your money), and that die as soon as the sunlight of security or complexity hits them.
There has to be another way. Nobody questions the versatility LLMs provide, or the boldness they give you to take on new projects. But I am not convinced that programmers or software engineers are going to disappear. I am not convinced that a manager can ship a new feature in a couple of afternoons and then lay off, or skip hiring, a junior or senior developer. There is a line near the beginning of Dune that goes something like this: "there was a time when men handed their thinking over to machines in the hope of being freed, but that only allowed other men with machines to enslave them."
We need to find a new path between smartphone addiction and being dumbed down by LLMs.
Of course it is. The AI coded apps only (debatably) work as end-user apps. They are thin little bumps on top of the software mountain. So much, including all the AI tools, depends on efficient, perfectly deterministic tooling.
It is almost hypocritical for AI enthusiasts to say we don't need to look at code anymore, while everything they do is build on piles and piles of infrastructure and tools that are carefully hard-coded, deterministic building blocks. The whole thing would come crashing down if everyone embraced vibe coding at every level.
I am just constantly baffled by people thinking that because we can make (crappy) end-user apps and one-off tools, no one needs to code anymore. Does anyone build solid tools and libraries? Does anyone still get pleasure from building a rock solid, snappy app?
I can feel all the AI coded apps and website turning to ever buggier sludge under my fingertips, and people are rejoicing that this is the way of the future.
I refuse. I will not touch Claude.
If you want to help Linux and are an artist, I plan make the initial source code release with 3 different tables and a guide for how to create your own pinball tables. I am looking for artwork to help with table artwork. My contact information is on the page linked to this submission.
Miss indicated that a woman was unmarried. Mrs indicated that she was. Ms was added to behave like Mr in that it doesn't indicate marital status.
My reasoning was that that he had become too powerful while largely avoiding politically charged positions, or other "stuff" that can be used to "control" him.
The political and intelligence establishment doesnt particularly like powerful individuals who remain politically independent.... outside of their control.
Part of the point of all this is that he's supposed to be in control of his emotions and able to make the best decisions without letting his emotions interfere.
If he can't even make the very simple decision to keep his dick in his pants in this very clear breach of commitments he has publicly made, then what other terrible decisions is he making? What other commitments will he break because it feels good and he can't control himself?
But the extramarital affair with Alice Huang is quite different. NYTimes reports that Alice Huang is a Chinese spy, with whom Jensen Huang may have conspired to break the law, selling advanced NVIDIA chips to China. That affair puts the whole company at extreme legal risk. Gates, Bezos, and Brin have nothing like this on their records that I'm aware of.
As I write this, the top post on HN speculates that this is "the tip of the iceberg," indicating that there at least may be more affairs yet to be uncovered, and, plausibly, Jensen Huang may have committed even more enormous mistakes with serious consequences for NVIDIA.
“It is generally considered polite to address a woman by Ms. rather than Mrs., unless the preference of the woman in question is clearly known. This is especially true in written communication, as dictated by professional etiquette. […] It is normally considered correct to address a woman as Ms. [Lastname], regardless of her marital status.”
certainly sounds like that's the case, yup
Many people find it demeaning to address women differently depending on whether or not they are married, especially when we don't do the same for men.
Dumb language question: if Powell is married, why isn’t she referred to as Mrs. Powell in this article?
In addition Nvidia is incredibly tech-heavy company. You need another EE with experience with AI chip design and business. Just having good experienced businessman looking at simplified powerpoints before making decisions of technical issues is probably not enough.
I was told later by a docent that my reaction was not uncommon. >smile<
The remainder of the valid Perl programs are mostly valid due to a Perl
language feature that is not commonly present in most programming languages.
Namely, Perl has a feature called “unquoted strings”, in which a sequence
of alphanumeric characters by itself is parsed as though it were a quoted string.
As an example, Figure 5 is read by OCR as the text ME, which evaluates to the
string “ME” even though the ME isn’t quoted. This would result in a syntax
error in most other programming languages.
(To be fair to Perl, when perl is run with the -w flag to enable warnings, it does
helpfully inform the user that at some point in the future, the Perl developers
will most likely pick gggijgziifiiffif as a new reserved word:
Unquoted string "gggijgziifiiffif" may clash
with future reserved word at - line 1.)EDIT: From their 2026 publication
>Permission to make digital or hard copies of portions of this work for personal use is granted; permission to make digital or hard copies of portions of this work for classroom use is also granted, but seems ill-advised
2024 (515 points, 125 comments) https://news.ycombinator.com/item?id=40197013
2021 (431 points, 165 comments) https://news.ycombinator.com/item?id=27929730
> Don’t use the emulator to test the following:
> - IAM policies and permissions boundaries
> - Resource-based policies, such as bucket policies
> - AWS KMS key policies
> - Access across AWS accounts or across AWS Regions
> - Any other security control
> Some services in the emulator evaluate IAM policies for IAM users and roles that you create in the emulator, but not the same way as AWS. The emulator never denies a request signed with its own key. It doesn’t check AWS KMS key policies or Amazon S3 bucket policies. It also doesn’t support calls across accounts or Regions. So a request that AWS denies might succeed in the emulator. A request that AWS allows might fail.
This is a shame because I'd love for the official AWS product to be better than e.g. ministack and localstack on this, which will go some of the distance. I think because the agents were only building on the public API surface, it's hard to build this part. But they should've tried and let the agents read whatever policy engines AWS has. Or maybe it's a way to prevent someone from trivially running aws-emulator as AWSaaS?
[0] https://docs.aws.amazon.com/emulator/latest/userguide/securi...
Google builds up a product over time and abandons it. This is annoying for long term adopters.
AWS has a habit of throwing a bunch of products at the wall, but only bothering to maintain & build up a few of them. This is annoying for early adopters.
That said, it’s been announced publicly here: https://www.aboutamazon.com/news/aws/how-autonomous-ai-agent...
Additionally Swami, our SVP of AI blogged about it separately and our CTO is tweeting about it - which is more than a number of products get in terms of coverage.
Floci[0] has been great for me. Use the Testcontainers[1], too.
Bonus: works for GCP and Azure, too (so far have only used in limited capacity, though).
Papers please has a much simple mental model.
Bo, this looks very cool, but I'd encourage you to immediately add "Based on Papers Please by Lucas Pope", with a link here:
Why not just sell to not-sanctioned entities, ensure compliance to the best of your abilities and otherwise keep your head down? Again, wasn't like they struggled to find buyers for their chips.
Considering Megaspeed, I now see this in a very different light. Him cheating is personal and I wish the affected members of any family of a cheater the best, but not relevant for us outsiders.
Being honey trapped very bad securities wise, but he'd be a victim (and evidence for why we shouldn't discount Mistral in the frontier lab race [1]). If sanctions avoidance went all the way to the top, that is hard to negate even for a board that is interested in keeping him.
[0] https://www.youtube.com/watch?v=1H3xQaf7BFI
[1] https://time.com/archive/7146565/hollandes-popularity-goes-u...
> If a post reports on something found on another site, submit the latter.
- They only lock normal action calls. If your workflows invoke other reusable workflows, you need to keep those hash-pinned manually for now.
- Composite actions are not themselves locked. Instead, they get locked transitively through their callsites i.e. the workflows that call them. This makes any individual caller that uses lockfiles safe, but it also makes the current `gh actions-lock` default unsafe (since it removes hash-pins from the composite action itself, meaning that any external caller that invokes the action loses their transitive pinning).
- Other forms of `uses:` clauses, like Docker images, are also not currently locked via the lockfile. And the same for forms of interior mutability (like an action running `npx ...` under the hood, although the latter is fully outside of GitHub's control to prevent. But these mean that you can't assume that a fully locked Actions setup is in fact 100% immutable.
I hope they address these gaps soon, because it's otherwise an excellent (and long-awaited) feature.
A tragic loss of consumer surplus and incredibly inefficient in terms of resources, bandwidth and people’s time.
It’s sad but it seems inevitable at this stage.
Condescending attitude aside, if I wanted to hear what Fable says about this, I'd ask Fable myself. I want to hear what YOU have to say about this.
Clues specially being
1) you hope to RETIRE
Ie be able to not work
2) you HOPE to retire
Ie it’s not certain
1. Boyce, Brown & Moore (2010), “Money and Happiness: Rank of Income, Not Income, Affects Life Satisfaction”
We need to react to it with the same intentionality that we did a century ago when we won the 40 hour work week, which still hasn't been entirely eroded despite half a century of chipping away.
A typical tech job requires far more than 40 hour work weeks.
Someone hit on the notion that it’s Stockholm Syndrome. You bond with people you go through shit with. You don’t bond with people who prevent the shit. That was a super awkward conversation. The asshole in charge quit shortly after it, but 2 years later we still occasionally called each other out for making plans that accounted for a toxic chain of command that was no longer toxic.
Exactly why the general public thinks Y2K was a "hoax" or "overblown".
That's a common bias in management. Heroic fire fighting gets rewarded while silently preventing a crisis with proper planning isn't. Nobody ever gets credit for fixing problems that never happened.
This anecdote hits differently for me. I grew up reading about Fog Creek and consuming all of their writings (and that one film) as an example of what a software development company should look like. It feels strange to have the curtain peeled back to reveal that something like backups, which would be considered basic requirements today, were not part of their early operations.
It's not a mystery, it's exploitation.
With all the business efficiencies we've made with computers as devs and IT, we should all be working 3 days weeks, instead a few people are so rich they can buy elections.
Context: in 2012, I literally quit Fog Creek, largely over the fact that we had no backup strategy and I was told that a backup strategy was not feasible. I was "convinced" (paid a lot) to stay on for another two years to see if Kiln would succeed. On my way out of that meeting, I commented to a coworker, "I mean, what's the chance a hurricane hits Manhattan in the next 24 months?"
100%. The chance was 100%. And the only reason people had to haul open containers of diesel up 17 flights of stairs was because no one wanted to admit that the chance wasn't 0%.
Anyway, not sure what was going on with Jane Street here, but holding up that Fog Creek/Trello/SquareSpace story as a positive always drives me somewhat nuts. It lionizes a failure.
I had made sure to observe the filter replacement process in case we needed to do it again ourselves. We did, and the next time someone recorded me doing it so everyone could learn and we handled most I’d the rest of the swaps from then on.
There were a lot of memorable experiences that week or so.
It took me an embarrassingly long time to realize that (my 30's), and now I keep my conversations with my coworkers very utilitarian and divulge minimal information about myself. I try not to be a dick or anything, I'm polite enough, but pretty much everything I talk about is either directly relevant to work and/or so surface level as to border on "vapid".
ETA:
Just as an example, I used to have really bad sleep issues in my 20's that would make it so I wouldn't be able to get to sleep until around 4am, if at all. I made the mistake of bringing that up at work to my manager. Two weeks later I got a phone call at 3am because there was something he wanted to look at and "he knew I would be up". I was up, but I didn't really have much desire to get my laptop out and do work. I liked that job at the time and I liked my manager, so I did it anyway, and it wasn't until years later did I think about that incident and how utterly weird it was that I was sort of being punished for my insomnia.
My sleep issues have gotten a lot better now, so if someone calls me at 3am I probably won't hear it because I'll be asleep, but I remember that incident now and have it constantly spinning in my brain when I chat with coworkers.
“The superhuman folks at the data center are hauling 55 gallon drums of diesel fuel up 17 flights of stairs. Assuming they're able to sustain a rate of transfer that exceeds the rate of consumption, the data center should run until electricity is restored. The latest estimate from Con Edison for lower Manhattan is 4 days. (The New York Stock Exchange will be open tomorrow, also running on generator power.)”
https://news.ycombinator.com/item?id=4720388 (https://web.archive.org/web/20121101194314/http://status.fog...)
https://arstechnica.com/information-technology/2012/11/how-o...
Please, it is still a research project, it's not like a regular computer just faster, it is novel research.
Regarding the quantum side, I would love to know if by programmable qubit are they referring to an error corrected qubit or just one where gates can be applied.
I suspect it means error corrected qubits. I was wondering if there was work being done on algorithms that are noise resistant and as such can take advantage of the raw qubit count instead of the error corrected qubit count.
I would love to work in this area but I think it's all research at the moment. It's really interesting, how it actually works, not just the hand wavy abstractions.
Jensen banged a customer and Chinese spy while advocating for lifting exports controls on China and corruptly quashing investigations into her firm [1].
He couldn’t have turned this into more of a public spectacle and concern if he tried. (Coming clean right after the Congress flipped didn’t help.)
[1] https://www.nytimes.com/2026/11/09/business/economy/nvidia-m...
But that is not the case. He is the founder. A founder has a very different relationship to a company. He is its father and mother, and severing that relationship and transitioning to a CEO that does not have the same relationship is fraught with danger.
As a shareholder, for purely self interested reasons, I would be very hesitant to call for his removal.
doesn't sound like valid reason
From my own experience, these learned methods from Apple, Spotify and others, are highly relevant to hyperscale only - when you are dealing with hundreds of millions of items to choose from (or more), and require super high relevance in the top-5. For all other use cases (up to a few million items) I found that just building a BM25 index, and giving the LLM a BM25 "search tool" and casting a "wide net" wins every single time compared to learned methods (though the latter are more publishable in recsys and icml).
[1] https://research.atspotify.com/2025/9/you-say-search-i-say-r...
We have made a determination that more hype is needed
My goal was not to imply that we're not shipping faster than we were prior to AI, as I think that's beyond debate at this point. What we investigated is why we're not shipping faster than we are today. If (high quality) code is being written at five times the rate it was before and we're shipping "only" twice as fast, then what's the reason for the difference? The talk shares results from studies that help improve our understanding of why.
However, for those that are actually curious, the submission is not itself a cynical or dismissive take. The question in the title is interrogative and contextual, not rhetorical and universal, and the presentation tries to break down some of the bottlenecks and challenges that are leading to AI integration having a different impact on software engineering output than some people might expect.
Rather than suggesting those bottlenecks/challenges are universal or intractable, it just enumerates some of them and presents some survey analyses on how widely encountered they are.
It's actually pretty interesting and the kind of thing that engineers and those who work with them want to be reflecting on, so that they can better engineer their usage and processes around AI itself.
"One occasionally reads newspaper accounts of how two programmers in a remodeled garage have built an important program that surpasses the best efforts of large teams. And every programmer is prepared to believe such tales, for he knows that he could build any program much faster than the 1000 statements/year reported for industrial teams. Why then have not all industrial programming teams been replaced by dedicated garage duos? One must look at what is being produced."
Brooks argues that the coding that happened in the garage is only a fraction of what you need for the final product:
"It costs nine times as much. But it is the truly useful object, the intended product of most system programming efforts."
What we're seeing right now does not feel much different. You can greatly accelerate a first draft, a greenfield project, an experiment, or at least the initial iteration of one. Changing a live production system or a widely used library is an entirely different proposition.
edit: if you're seeing a slow down in PRs, it's more likely that 10 of your 10x'ers are slinging code in 3-4 plus accounts in codex on the side during their shift on 8 startups.