You never know when it's going to vanish
Source: https://opensourcemalware.com/blog/agent-framework-pypi-comp...
One thing I still regularly use jQuery for, though, that's rarely mentioned, is its event model. You can put a listener on any object, not just things in the DOM. Any data object can trigger an event. This is incredibly powerful for isolating logic from DOM elements and events. A frequent use case is for DOM elements to contain a reference to a data object that may be replicated elsewhere on the page. The data object sets up and tears down visual elements, listening for interaction, then dispatches a custom event. Your core logic listens for events on the data object instead of on every individual element that might refer to it. I usually delegate the draw routine for e.g. html snippets of a card or an interactive element to the data object itself in a draw() function that returns a string which incidentally jQuery handles quite nicely inserting into the DOM without needing to explicitly call createElement.
I wrote my own implementation of this a long time ago, but in the end, that feature plus the other ergonomics of jQuery still make it my go-to today over vanilla js. And pretty much everything I write is without any reactive framework (other than my own vanilla component/templating/remoting system), so the overhead is barely noticeable.
That was, and still is, rare! And in a time without type checking and bundling, it was vital. Match your parentheses and quotes, and your jquery code might not find what you want, but it would (likely) not throw exceptions.
For those unfamiliar, I recommend looking at https://github.com/DefinitelyTyped/DefinitelyTyped/blob/mast... and searching for the 201 instances of
: this
in that file. A remarkably fluent API. We could all be so elegant.And while most of the use cases for jQuery have long since died out with things like classList, querySelector, etc, I do feel there's probably value in a framework that focuses on minor enhancements to a site rather than being the backbone of the whole thing. Like if you want a React or Angular or Vue equivalent, something like jQuery would be a terrible choice, but a lot of sites are static ones where the only scripting needed would be for minor things like dropdown menus or slideshows, and I can see a modern, lean library like this being a decent choice there.
jQuery made the decision to do this, not JavaScript.
There's a list of large legacy sites that use jquery behind that link, including:
* Microsoft.com * Pinterest.com * Adobe.com * Yahoo.com * Wordpress.org * Digicert.com * Samsung.com * Archive.org
I switched it to that model and it was able to resolve tricky python build issues within Windows WSL/ubuntu and also generated a good number of test cases and ensured all of them pass.
It resolved all the Azure deploy (bicep) issues as well. Didn't feel the difference from Claude Opus.
Microsoft didn't win spreadsheets by having the most advanced or releasing the most often. They made a really good, easy to use product, and got it in front of everyone. (Though Office's p-code was a huge advantage to portability)
Anyone has more background how they were able to catch up that fast? Distillation of GPT due to their OpenAI acess to the model might make most sense but they probably are not allowed to do that.
That is a very tenuous lead and doesn't speak to much of a sustainable competitive moat for either company. The have to keep sprinting as fast as possible on the hamster wheel to just stay ahead.
1. The CS-3 operates on average at <65c, well below its thermal limits. Considering the actual thermal parameters of the wafer, using the short seller's own temperature equation would imply the wafer can operate at 85c for 33 years
2. "Ghost tokens" are not unique to Cerebras and the 3x lifespan reduction claim is based on extrapolating GPU figures which is a big logical leap
3. "since the introduction of the CS-3 in 2024, less than half of one percent of all CS-3 wafers deployed have been decommissioned."
https://cerebras.ai/blog/what-it-took-to-build-wafer-scale-a...
Disclaimer: I work at Cerebras
You're being extremely generous if you're reading that article and coming away with any sentiment except "this looks suspicious and shady".
The article itself is extremely clear on the timeline being suspicious.
> As recently as the summer of last year, Singapore was officially one of Nvidia’s NVDA 2.87%increase; up pointing triangle biggest markets, generating 22% of the chip maker’s revenue. Then, practically overnight, it disappeared from the company’s financial disclosures.
> Investors deserve to know why, and Nvidia needs to start explaining, with or without Jensen Huang at the helm.
Then a few paragraphs down:
> The next fiscal quarter, which ended in October 2025, Nvidia switched reporting methods. It began reporting geographic revenue “based upon the location of the customers’ headquarters,” rather than “customer billing location” as it previously had done. The change came amid intense scrutiny from the U.S. government over whether restricted artificial-intelligence chips were being routed to China through overseas intermediaries
Regarding your point that they didn't just "stop reporting" the revenue but rather "stopped reporting it as coming from Singapore" - isn't that a given for anyone with basic financial literacy? Not reporting revenue is illegal. Shuffling the books to reduce scrutiny on where it's coming from is legal but shady.
They didn't stop reporting Singapore revenue, they made a change to reporting such that revenue that was being reported under Singapore ends up reported under US, China and Taiwan revenue. This was reported by wsj a week ago. Here is a better article
https://www.wsj.com/business/earnings/nvidias-disappearing-s...
As far as I can tell this FT article has no new info in it
https://www.irishexaminer.com/business/companies/arid-418199...
https://en.wikipedia.org/wiki/Apple%E2%80%93EU_tax_dispute
Apple and Singapore, https://finance.yahoo.com/economy/policy/articles/singapore-...
https://www.irishtimes.com/business/2026/11/08/singapore-ove...
The 104B number is the external revenue of NVIDIA's Singaporean subsidiary in 2025. This includes any orders for GPU chips that happen to get routed through the Singapore entity. For those unaware, end customers generally don't buy anything from NVIDIA. Customers by NVIDIA GPUs through a large network of OEMs/ODMs and resellers. Many of the largest companies in this network are based in east asia (eg. Wistron, Foxconn, Asus, Lenovo) and when your local hyperscaler makes a server order, one of these OEMs places an order to NVIDIA for GPUs. Because of this much of this revenue likely comes from end customers all over the world, including the US.
In fact, in one of the 2025 earnings calls NVIDIA stated that 99% of the singapore entity's revenue for export-restricted chips came from the US.
It's a little bit difficult to untangle exactly how much of NVIDIA's revenue came from end customers in Singapore in any given year. In 2025 it looks to be about 19B. That might be unreasonable for a country singapore's size, but it's certainly not half of NVIDIA's revenue.
The 104B number is NVIDIA's Singapore subsidiary's total revenue. This is just any order that gets routed through the Singapore entitiy regardless of where the chips end up, and because many of the server OEMs that actually sell GPUs to end customers are located in East Asia (eg. Wistron, Asus, Lenovo, Foxconn) it's likely that much of this revenue results in chips being shipped all over the world, including the US.
It's also a massive tax haven, so there's that.
This is an accounting question of how their revenue is attributed. It means nothing about where the revenue is generated or where the chips end up or to which subsidiary they are sold to. If you were trying to hide the revenue from regulators, this is not how you would do it.
Even IF this decision was in response to foreign export controls being put into place, it's just as strong an indication that they were planning on making less money via Singapore and didn't want to show that to investors.
So maybe she’ll be gone next once the new CEO comes and cleans house.
I dont believe the board fired the founder and only CEO they ever had over a series of affairs. They fired him because one of the affairs was with a lady pretty clearly helping import chips from Singapore to China.
Wth? This is a public company, how can they hide 100B in revenue?
https://en.wikipedia.org/wiki/Soviet_Project_K_nuclear_tests
>The electromagnetic pulse (EMP) fused all of the 570-kilometer monitored overhead telephone line with measured currents of 1500 to 3400 amperes during the 22 October 1962 test. [...] The EMP from the same test caused the destruction of the Karaganda power plant, and shut down 1,000 km (620 mi) of shallow-buried power cables between Astana (then called Aqmola) and Almaty (then called Alma-Ata).[3]
I would be more interested in a detailed military tech case study in the real world, or a literary analysis of Frieren, both of which I'm sure already exist.
Edit: I also would like to be given paragraphs rather than having to read one word at a time and sit through a transition scene each time I try to read the next.
Edit 2: I gave up reading it halfway through because scrolling too fast ran out of memory and crashed my (mobile) browser.
https://en.wikipedia.org/w/index.php?title=User_talk:%C3%8Fv...
> The Arbitration Committee has reviewed a dossier of "Tech4Palestine" Discord server related evidence and has determined that, as of this time, the concerns raised have been adequately addressed. The evidence has been retained by the Committee to be used, if necessary, to corroborate additional evidence received. The editors who were brought up in the report are:
* Samisawtak (talk · contribs) - topic banned by ArbCom
* Ïvana (talk · contribs) - banned by ArbCom
* Tashmetu (talk · contribs) - EC revoked, warned for violating ECR
* Isoceles-sai (talk · contribs) - Result of below
* ZionistHater (talk · contribs) - Blocked as WP:NOTHERE by Drmies
* Discourseofcourse (talk · contribs) - Blocked as WP:NOTHERE by Bbb23
It's telling they refute the ADL report (which is non-credible, written after the bans, and had no influence on this) and don't mention any of the other reports describing what they actually did.
The consequences of such biased editing (on both sides) are widespread, especially since LLMs now rely heavily on Wikipedia as an unbiased source.
[1] https://www.adl.org/resources/report/damage-done-wikipedia-l...
Maybe that's just marketing, but it's really intriguing how human best practices are also agent best practices; how, maybe, we've reached a point where even devs aren't recognizing agent best practices as human best practices anymore; and how suddenly best practices are being prioritized in a way they often weren't - humans could/would/were expected to suck it up and deal with poor devex, but agents simply can't. A strange sort of victory for us folks who cared about it.
And of course the phone itself had major issues, it was overpriced and under-featured and they never made good on many of their technical promises either. The whole experience left such a bad taste in my mouth that my Net Promoter Score for them is below zero. I only regret that I didn't make it to this thread fast enough to tell people to stay away.
This is a phone that respects the user - not Samsung, Apple, Google, etc.
Reading all of the negativity, it sounds like people are implying we should never try because the first or second attempt will never have secondary features as goof as the mass market phones.
Remember: the primary feature is a phone that you truely own and can trust - you can't do that with the others.
Just read this article this morning which I think highlights the problem: https://www.newsweek.com/cyber-attack-websites-affected-1272...
I think it's really cool that browsers are VMs that can run Doom now but I really think we need to come back and focus on the basics again: HTML as hand-editable documents you read, more interface components, accessibility tools. Not sure the Gemini protocol (https://en.wikipedia.org/wiki/Gemini_(protocol) will succeed but I understand and loved the spirit of it.
If it was AWS fine, that’s not uncommon, but there have been several outages, including at the nsa, which is also unusual.
The only chance we have for a more reliable world is if these outages and hacks become more and more common. Institutions and people will be forced to wake up and find more local and diverse alternatives...
Edit: and apparently Apple takes appointments but has no methodology to cancel them due to catastrophic outages like this, nor do they have any ability to conduct non-digital transactions…typical tech
AI is tuned to speak confidently and ship independently; it is rare that it will suggest that one go back to a broad survey of stakeholders to ensure no Chesterton's Fence is being torn down (https://fs.blog/chestertons-fence/). "I've read the documentation and logs, and nothing suggests this system needs these legacy capabilities/components/capacity" is something we hear all day, and it is a siren's song if there ever was one.
Couple that with a general (and valid!) consensus that modernization of legacy systems must be accelerated for a variety of reasons - to mitigate this new and continuing "Black October" of hyper-agile agent-driven cyberattacks (whether the network is public or private); to have more flexible infrastructure for new features to stay ahead; to allow human teams to show their efficacy in a world where technology jobs are increasingly at risk - and you have a perfect storm where change management is no longer what it once was.
- stable always-on service that never goes down
And
- down 1% of the time, but FDIC insured store of assets
I know which of the two I’d choose amidst an unparalleled level of security risk
> Over the past week, I’ve been overwhelmed by the love and support of friends, industry colleagues, and people I've never met. Thank you to everyone who has shown kindness to my family and me. It means the world to me.
> I came to the United States 53 years ago, and becoming an American citizen is one of the greatest gifts and proudest moments of my life. NVIDIA has been my life’s work. Building it into the company it is today alongside my colleagues has been a privilege and an honor. Nowhere but in America could a company like NVIDIA be possible. I love our country, and I love NVIDIA.
> I did not live up to the standards expected of everyone at NVIDIA, and I have made serious mistakes in my personal life. I created problems for the Board that I deeply regret. I own my mistakes completely, and I am committed to cooperating with the Board’s independent investigation.
> Since my departure, I’ve seen an outpouring of calls from NVIDIA’s employees, investors, business partners, and customers for me to return as CEO. Many continue to reach out to me directly as the company navigates complex technology, market, competitive, and geopolitical dynamics. NVIDIA is a once-in-a-generation company. If the Board of Directors decides that my return is in the best interest of our shareholders, our employees, the company, and America, I would be honored to answer these calls and return as CEO.
> Thank you again to everyone who has shared their good wishes. I am profoundly humbled by your support. To my family: I love you, and I will spend the rest of my life being worthy of yours.
> Jensen
For all of their flaws, I think universities are one of our greatest inventions and it's a mistake to view them purely in economic/transactional terms.
1. LLM performance has more or less peaked. Since Fable 5 released in June, every Claude model has been a ~1-2% improvement over the last one (if you go by benchmarks). Fable 5.2 marketing verbiage has shifted to how it's "lower cost per task" which imo marks a regime shift. I would bet money that this time next year, we're not going to have a model that's obviously and significantly better than the best ones we have today.
2. If you hold that this is about as good as LLMs are gonna get, we haven't found ourselves at a point where we can completely remove the software engineer from the seat. A software engineer is still required to create real commercial-quality software. The role hasn't been commodified either - a better software engineer still writes better software.
I don't get why people keep repeating this as if we have any idea. How do you know AI will be a power tool to those that know the fundamentals? How do you know it won't replace those people anyway? Why is knowing the fundamentals of computer science going to be the thing that keeps people from getting replaced by AI? You're just repeating something that sounds catchy, as if that somehow makes it more true.
Are requests for courses being accepted? :)
If yes, may I humbly request Larry Wasserman's Intermediate Statistics, Dave Andersen's Distributed Systems and Dave Eckhardt's Operating Systems (or whoever is teaching it now).
If you learn the fundamentals properly, then AI will be a power tool for you. If you don’t, then it will replace you.
I don’t mean this in an aggressive “you better get used to it” sense, I realize there’s nuance - but this is the version that fits on a T-shirt.
I plan to put my distributed systems lectures online the next time I teach it, but not sure when that will be after I rotate out of teaching 15-213. One of my colleagues who also teaches it might do it before then.
We're releasing full semesters of lectures from current CMU computer science courses across a variety of topics, with one course premiering each day this week. More courses come in January.
Yes, MIT got here 25 years ago :). What I'm excited about is how we're doing it: a few of our students built tools that automate a lot of the production work, so a small team can get a course out quickly and still have people review every lecture.
We'll have lots more coming across a wide array of subjects. I'd love to know if there are topics you would like to see us cover!
The problem is, to justify the substantial investment in data centers and the eye-watering valuations of Anthropic and OpenAI, AI has to successfully reach into essentially every niche of the economy. It can't just be a tool to code faster. It must make most jobs faster period. From daycare to farming, we need relatively short-term ways of integrating this technology.
AI will survive the hype cycle, but the AI companies very well may end up like railroad companies or Iridium, where the shareholders got zeroed out and then someone comes in, buys the assets for 10 cents on the dollar, and builds something useful out of it.
Over the past week, I’ve been overwhelmed by the love and support of friends, industry colleagues, and people I've never met. Thank you to everyone who has shown kindness to my family and me. It means the world to me.
I came to the United States 53 years ago, and becoming an American citizen is one of the greatest gifts and proudest moments of my life. NVIDIA has been my life’s work. Building it into the company it is today alongside my colleagues has been a privilege and an honor. Nowhere but in America could a company like NVIDIA be possible. I love our country, and I love NVIDIA.
I did not live up to the standards expected of everyone at NVIDIA, and I have made serious mistakes in my personal life. I created problems for the Board that I deeply regret. I own my mistakes completely, and I am committed to cooperating with the Board’s independent investigation.
Since my departure, I’ve seen an outpouring of calls from NVIDIA’s employees, investors, business partners, and customers for me to return as CEO. Many continue to reach out to me directly as the company navigates complex technology, market, competitive, and geopolitical dynamics. NVIDIA is a once-in-a-generation company. If the Board of Directors decides that my return is in the best interest of our shareholders, our employees, the company, and America, I would be honored to answer these calls and return as CEO.
Thank you again to everyone who has shared their good wishes. I am profoundly humbled by your support. To my family: I love you, and I will spend the rest of my life being worthy of yours.
Jensen
Wow! I love this.
I love that you bothered to email them with a thought out question.
I love that they bothered to respond with a thought out answer and they took you quite seriously.
I love that you bothered to tell us about the exchange.
Thank you so much for all of this. This is what we should all strive to do with our time and as a community.
Compare with other causes of death for kids
- 29 from vehicular heatstroke
- ~300 (1/100k) familial homicides
- 1014 traffic fatalities
Even if it were a concern having a phone in such a scenario may make for good hollywood, but is terrible from a safety viewpoint. The last thing you want during a lockdown is phones clogging the network and ringing as kids try to hide.
My message to them:
Dear Dr. Burland and Professor Kane,
I was reading your November 2026 report, Phone-Free Schools and Achievement Growth, and saw that the study was preregistered on the Open Science Framework. I haven’t been able to locate the registration.
Could you please share a link to the preregistration or let me know where I can find it? I’m particularly interested in checking whether the exclusion of extreme fall-to-spring score changes was specified in the original analysis plan.
Thank you for your time and for making the research available.
Best regards
Their reply: Rahim: That part of the study was not pre-registered. We have used the NWEA data from earlier years, but had not encountered such extreme outliers. The magnitude of gains or losses we dropped were clearly errors—multiple standard deviation gains in a single year, also inconsistent with the midyear measurement. I’m ccing our colleague, Samantha Cullum, who can provide the link to the pre-registration. Tom KaneSo the effect could be tiny.
Worse, the report says it excludes observations outside the 0.5th and 99.5th percentiles of the 2021–22 fall-to-spring score-change distribution. I'm not sure how this is defensible. Either you trust your measurement instrument or you don't.
If you're using a difference-in-differences design, excluding observations can change what you're measuring, particularly if the exclusions differ between treatment and control groups or over time. Without trimming those observations, the estimated effect drops from 0.0087 SD to 0.0058 SD (and is no longer statistically significant).
The authors preregistered their study, but there's nothing in the report indicating that the trimming of 'extreme' values was part of that preregistration...
China just banned them outright. It's the same with companies - weak committee vs. a founder who just makes the hard call.
It appears to be either:
A. An excuse to get their kid out of trouble for scrolling TikTok & Reels during class
B. A weak attempt to provide a socially-justifiable reason for the parent’s general anxiety (constantly texting their kid, watching their precise location, using Life360 to see their messages with peers and fine-grained activity)
C. A futile effort to exercise control over a hypothetical situation where the parent is stripped of control. What are you going to do if you get a text saying “there’s a shooter in the school”? Call the police? The faculty all have phones and are doing the same. Text them back saying “okay I love you”?
Yes, it is important to understand which things are true and which are false.
Might as well leave your kid home at that point, iPad in hand. Even less ""risk"" when they were never going to learn anything at school anyway if they are allowed access to devices.
The problem with phones, as well as social media, or anything else with a network effect, is that it becomes a prisoner's dilemma. You can recognise the harms, but feel the need to take part anyway, because opting out puts you at a social disadvantage. If everyone opts out, the disadvantage is neutralised
I don't understand when it become allowed honestly. Like when I was at school, phones were never allowed. If a teacher saw one they'd confiscate it and your parents would have to come and collect it.
That was before the absolute worst addictive dark patterns etc. Which means that someone looked at the world after that point and said "yeah actually, phones are ok now" which baffles me
Bonkers to me that we ever allowed this.
Oh, and for context, the kids are allowed to go to the office and check in on a message or send a text (e.g., soccer practice is canceled due to rain, can you pick me up at 4?) during any non-class period so it's still easy to communicate.
A serious effort here would use a specific set of terminal bench questions. Instead these look like cherry picked questions. How were they picked and why? Without an answer we should assume they were picked to produce some result.
It’s unclear whether that’s supposed to be surprising against the selected models for the benchmark. I suspect it’s not and the headline is written to pander to “wasting money on AI” upvotes.
Edit: they changed the title. The one used to get to the front page was something like "Across N agent runs 64% of agent cost went to failing runs".
Edit2: since the link is poor quality, I tried to answer some of these questions myself using 927k agent runs vs. this blog post's 27k: https://micjm.com/blog/on-model-harness-interactions/. The upshot: there's not a lot of harness x model interactions. OpenAI models do seem to have a preference.
> Don't post generated text or AI-edited text. HN is for conversation between humans.
Blazing fast compiler (on that era's hardware), glorious TUI framework (Turbo Vision) and a nice help system. Anders Hejlsberg is a genius.
After using this and Delphi and then having to develop UIs on the web it was like losing an arm.
It’s not a great loophole per se but it’s a flawless example of what I might call “loophole thinking” which is kind of a transition from individualism to selfishness that has been happening for a long time (maybe there was never much of a difference)?
"The average Japanese fears not the judgment of an omniscient God, but the judgment of his or her fellow Japanese."
Loopholes are social slack. You can speed to the hospital, usually, in an emergency. If you’re running late to pick up your kid, sure, double park or streetside with hazards.
But if you turn this into a lifestyle, you’re broken. (Every person tweeting with religious zeal for Jensen falls into that camp.)
Here's my most-recent real-world problem: I wanted to plug in my midi controller keyboard and just play a piano, so I can practice my choral music. Very basic.
This turned into many hours over several days trying multiple DAWs, trying to get tehm to install, get them to run at all, get them to recognize my controller, get them to play a VST plugin...changing system settings, completely jacking up my sound, restoring it all again.
It was a nightmare.
So instead I fired up Claude, worked with it for about 3 hours, and now I have an *8-channel* VST player that supports all my hardware, and includes a snappy UI, with outstanding performance and sound. It's amazing!
I literally spent less time creating a custom solution than I did trying to get the off-the-shelf stuff to work. I could make my player OSS too, but the effort to do that is huge, for all the reasons this forum would understand.
Bespoke software avoids all of that and allows me to do what I want, with the environment I have. I don't have all the features of an FL Studio, but I have what I need, and that works for me, beautifully, at low cost and effort. What's not to like?
So maybe we should share OSS "templates" instead of full products, projects that are the barebones of a solution that you can clone, and have your LLM modify to suit your needs, on your platform of choice.
For example, perhaps my barebones VST player is mostly good enough for you, but you also want to play sound fonts? Cool, clone my project, instruct your LLM to add sound font support, and add support for your controller or your output device. Or convert it from Gnome to KDE.
The time and cost of doing this is so little now, why not? I'm getting more out of my computer than ever, and I'm happy about that :)
Have you considered that the person submitting the PR to fix said problem has themselves experienced it? I find that is most often the case. But there is definitely an increase in the "performance edge cases, or hypotheticals" framing, where I am equally befuddled/too tired.
My open source projects were barely sustainable when the few PRs that came in would fix actual, non-hypothetical, serious logic problems.
Now I've been seeing PRs like that, fixing annoyances, performance edge cases, or hypotheticals, across my open source repositories, and I don't have time for them. I especially don't have time to fix problems that probably nobody will ever experience, and yet they keep coming. I have not organized my life around triaging annoyances (let alone fixing them).
If the trend continues, I'll soon be forced to close repositories to contributions, or just stop maintaining the projects entirely. And I imagine that's the point a lot of maintainers are at.
This reminds me of when "The Joel Test" was first published and it was met by people unbelieving there was a world out there who weren't even using source control at the time, let alone a bug-tracker.
A key pillar was introducing things like static analysis to provide cheap deterministic checks as the first goal and only then moving to other forms of verification.
There’s a long section in the middle where the article describes per-change verification via skill files and “ask Claude to use the chrome-for-Claude tool” when checking if changes have introduced regressions.
But there’s already tools which do this without spending tokens tokens. Playwright will achieve equal quality but deterministic results across the same verification attempts. It is write-once-use-often, so has a marginal cost near zero.
Im already spending my $2k usd per month in tokens with just one or two agents. How do they expect me to do that?
Also... glutaraldehyde is not keeping the bacteria alive. It's a biocide. but it does keep the structure intact as a fixative, pretty big difference:)
[0] https://socket.dev/blog/agent-framework-pypi-package-comprom...
[1] https://github.com/microsoft/agent-framework