The basis of them stating this is the lack of a check for a formal 10b5-1 plan.
There are a variety of other ways that he could schedule triggered sales that aren't immediately transparent to the media.
A lack of 10b5-1 does not directly correlate to not having any pre scheduled plan.
The fact that they're acquitting months to weeks, makes me feel like they're inventing news to throw on the pile of Nvidia headlines.
There are defined "open windows" he can sell, and these were sold in those permitted window.
It was also multiple batches over months.
He gave form 144 notice and portions were through trusts and charitable / foundation.
He filed multiple Form 144s notifying the SEC of proposed sales under Rule 144. Notable examples include a September 2026 notice for up to 5 million shares (roughly $1.09 billion) and earlier notices (e.g., involving 1.7 million shares in June 2026).
He's got this exact same pattern of behavior for the last 43 sales dating back to 2021.. it's the furthest that I left before I stopped.
That's the universal way to clean up roll back installs and other Windows update related.
If that's not working for you, there's probably a much larger issue at play on her machine and would be much better off backing up her data, wiping all the partitions and installing fresh. You'll likely reclaim more than 700 GB.
The only time I've seen situations like what you're describing is when these are machines that have been in use for years and have major Windows versions on top of major Windows version stacked on top of each other, not just incremental updates.
You don't have the option? Or when going into system mode as part of temporary cleanup it doesn't show older Windows update as an option to clean or it fails when it tries?
If that's important to you, there's an option to allow using a local account during setup by dropping to command line and setting a flag. Pretty straightforward if that's your only deal breaker.
I just recently installed Windows 11 on a computer and this is the steps that I followed
2) It opens a command prompt and you type OOBE\BYPASSNRO
When you hit enter it reboots and allows local mode.
Microsoft should not have hidden this away like they have and it should be an option for everyone installing Windows, and I recognize this isn't a solution for the non-technical user.. but for you specifically if it's something you're comfortable doing, it's an option to allow you to avoid having to use a Microsoft account when installing Windows.
I know that a lot of tech folks consider it a non issue because it is 'easy' to fix if you know the bypass and your way around the more inner workings, but broadly this isn't really an answer to the problem.
If we always just say 'Oh, well, if you do XYZ then you can manage to do A despite B' then we're eroding our control and agency. Slowly the bypasses will phase out, as they have slowly been in this case, and then we're really in a poor state.
I had a tech support client who came to be because they valued their privacy and didn't want an account but who had no idea how to make that happen. They didn't even really know what to be looking for.
Not only that, if someone has a tech knowledgable person set up their Windows 11 computer without online accounts MS updates will cause it to spam the tech-ignorant person owning that computer with misleading pop-ups saying they need to press this button to do $x thing. Which then secretly sets up online accounts which automatically and silently enables drive encryption. Which causes it to go to a bluescreen saying to go to some shady looking URL at aka.ms and type in a long code that appears on your blue screen only broken computer. It's basically ransomware. This is not some theory, I had this happen to a sewing machine laptop I set up for my mother.
And yes, I later learned that aka.ms is a real Microsoft URL. But when it was first described to me it sounded exactly like ransomware and had exactly that effect.
I'm by no means suggesting it's normal or that I think it's okay Microsoft removed the native ability.
You're right that I probably made an assumption that someone posting on hacker News would have the technical ability to hit a keyboard combination and type a command on a command line and hit enter, which is essentially what's required currently to utilize the local account ( https://rtech.support/guides/bypass-11/ )
My reply was intended as more of a direct message to OP, as opposed to general advice I think every end user should know and follow. I should have been more clear.
I would appreciate a link to the details about 'setting a flag.' The last version of that I used earlier this year stopped working in recent months, and I believe I saw tech news stories to the effect that Microsoft had intentionally closed that path.
For now I just use a customer-supplied email address which I used to setup a Microsoft account, setup Windows and use that account, immediately create a local-only 'admin' account with full rights, delete the online associated account, and start creating user account(s) from there.
I had recently had to install windows for the first time in years, did some chatting with mr gpt and poking around on reddit until I arrived at: https://massgrave.dev. They guide you on how to download windows LTSC from the "official" Microsoft links. So much less crap shoved down the system, it's kind of nice.
There are also community projects like NTLite ( https://ntlite.com/ ) and Windows ISO Debloater ( https://github.com/itsNileshHere/Windows-ISO-Debloater ) for the more technically advanced that want full control to prevent the bloat from even making it to your machine during installation at all (instead of having to clean it up after the fact)
If someone is already setting up Windows manually themselves, I believe they're already at some level of technical ability.. but as I pointed out in another comment, my reply was specifically to OP who I made an assumption had some technical ability given that they're here on hacker News. Maybe that assumption is wrong and I recognize that.
That said, the process is:
1) while installing Windows press shift f10
2) a command prompt will open and you type OOBE\BYPASSNRO
When you hit enter the computer will reboot and allow local mode.
I fully recognize that this is a hidden option that most people don't know about, but if you do, it seems pretty straightforward to me for someone that's already at a level where they're comfortable manually installing Windows, but I recognize I have an blind spot due to my own technical abilities.
Until they drop an update and retroactively force the creation of an online account. This is a game of cat and mouse that even the tech-savvy will eventually lose.
i couldnt get that to work the last time i tried but i ended up using rufus to burn the ISO which gives you options to turn off the microsoft account. it even disables the tpm2.0 check so you can install it on older hardware if you want to
If agents are differentiated solely by system instructions (persona and focus) while sharing identical models / similar model capabilities and contexts, I agree the architectural value is minimal.
However, a multi-agent system (MAS) utilizing true separation of concerns, where each agent is equipped with highly scoped tool access, isolated knowledge bases (RAG), and defined operational constraints, significantly optimizes accuracy, latency, and token consumption.
In the referenced example, the agents share identical capabilities, toolsets, and context windows. As a result, there is no functional role differentiation, rendering the multi-agent architecture redundant.
> Too much context confuses the LLM. Reduce unnecessary context and performance improves. Not to mention, too much context and you just run out of vram.
I don't think you can reduce context enough on a local LLM to get performance comparable to a 500k context without those reductions...
it seems to me that would only be true if the amount of knowledge needed to complete a task far exceeds the context limit. Why would that be the case? I find most people are using horribly inefficient contexts, and a naive approach to dark factories/software eng swarms seems to only compound this
>Whether it's the automatic charging cable snake, or your car picking you up and dropping you off at the airport, or renting out your car as a self driving ride share
While I can appreciate, and he has joked publicly about having "pathological optimism," that he is more aggressive with timelines than the reality of the alignment between technology, practicality (cost, etc.), and regulations, most of what you have listed continues to be progressively released.
Is he careful with his speech in terms of distinguishing a "big audacious goal" from a shipping reality in the near term? Absolutely not.
So going through some of your list..
>automatic charging snake (you left off hot swappable batteries btw)
So prototypes were made and demoed, but ultimately they landed with wireless charging pads.. they are already starting to deploy this for real world testing.
I'll preface this that I'm nearing 99% FSD on my Tesla, and most of that 1% was the early miles when I first got it.
Robotaxis exist and hundreds of them in multiple cities are autonomously driving passengers every day. They continue to scale those numbers monthly.
They also have 14 billion FSD miles and growing.
I use summon every day to have my Tesla unpark itself and pick me up at the front door of my office. Short of me holding down a "come to me" button on my phone, it performs this action fully autonomous.
Banish (the opposite that sends the car away) exists but is not publicly available due to safety, operational and regulatory reasons (for example where does the car go if it can't find a parking spot?)
Currently this tech is limited to line of sight (~200 Feet) but the speed and distance continues to increase over time with updates.. again as the tech and regulations come together. Eventually, as we see with Robotaxi, the ability to drive you to work, drop you off, go back home and park and return when you're off work is not science fiction. The end to end solution exists today, the technology and laws/regulation need to come together.
If your complaint is that cars purchased in 2016 can't do this, sure I'll concede.. but modem HW4 is more than capable and the tech and laws are continuing to adapt.
>renting out
A far future vision sold as near term, for sure.. but it's not a matter of if but when as tech and regulations close the gap. Robotaxis are essentially doing this today.
These are not "empty promises" more so than long term visions (that do manifest) sold as short term shippable options.
> These are not "empty promises" more so than long term visions (that do manifest) sold as short term shippable options.
By that logic any lie is simply a promise that hasn't been fulfilled yet.
> cars purchased in 2016 can't do this, sure I'll concede.
He made the promise in 2019! HW3 will almost certainly never have unsupervised FSD (let alone FSD without a human in the car), and HW4 didn't ship until what, 2023? There's more model years in existence where that promise is flatly broken than there are where it's maybe a possibility someday down the road.
But he also promised FSD rideshare in 2016, which arguably makes it even worse of a broken promise. Ten years later not only is it still a broken promise, they're actively working to ship a product that isn't what they promised.
>By that logic any lie is simply a promise that hasn't been fulfilled yet.
He has a long enough track record of actually delivering. Elon is not a new entity with an unknown track record. He's been active in running and growing companies for 31 years. He's been involved with Tesla for 22 years.
You're attributing statements as lies which I feel is too strong a term as it implies intent, which as far as I can tell, doesn't exist.
>He made the promise in 2019! HW3 will almost certainly never have unsupervised FSD
In the same vein of calling something a "lie", you anchor to another strong term "promise", which is also not applicable in this case.
What Elon actually said is that HW3 "had everything needed for Full Self-Driving"
In 2016, Tesla officially announced that all vehicles then in production had “the hardware needed for full self-driving capability at a safety level substantially greater than that of a human driver.”
Elon followed up that "All Tesla cars being produced right now have everything necessary for full self-driving. All you need to do is improve the software."
Both of those statements are true and HW3 is active with FSD today.
He went on record stating:
"Anyone who purchased full self-driving will get FSD computer upgrade for free. This is the only change between Autopilot HW2.5 & HW3. Going forward “HW3” will just be called FSD Computer, which is accurate. No change to vehicle sensors or wire harness needed. This is v important. "
He also honored this. Customers that purchased HW3 and the FSD package were provided upgrades when needed to support the modern 14.x FSD releases.
You're correct that they stated that HW3 would provide unsupervised self driving, when it was technically and legally possible and did publicly reverse course on that.
"Unfortunately, Hardware 3, I wish it were otherwise, but Hardware 3 simply does not have the capability to achieve unsupervised FSD. We did think at one point it would have that, but relative to Hardware 4 it has only 1/8th of the memory bandwidth of Hardware 4. And memory bandwidth is one of the key elements needed for unsupervised FSD."
This isn't a "lie" to plan for something and have those plans fall through, the intent was to deliver and it didn't happen.
He followed up that HW3 owners who bought FSD would get options like a discounted trade-in toward HW4/AI4 cars or a computer + camera retrofit (which would require special facilities). HW3 cars continue to receive supervised FSD updates (including a “Lite” version of later software), but unsupervised FSD is off the table without new hardware.
So providing monetary compensation in terms of trade in options or retrofits in support of transitioning them to unsupervised when available.
If we've already established that he makes ambitious goals and sometimes delivers late or pivots as needed due to technical or legal limitations, I'm not sure what you're on about.
Your tone suggests he is selling snake oil and never delivering, which just isn't the case. His companies have been years ahead of the competition in space and electric vehicle tech and self driving.
>But he also promised FSD rideshare in 2016, which arguably makes it even worse of a broken promise.
Saying something in 2016 does not make it a promise to deliver in 2016.
What he actually said:
"I think we are less than two years away from complete autonomy, safer than humans, but regulations should take at least another year."
He also described autonomous driving as "basically a solved problem."
Both those statements are true.. Tesla's are (and have been for years) safer than human drivers with FSD enabled, and those like me that are 99% FSD consider it full autonomous for all practical purposes. Yes I can't sleep in the car, but I get in, verbally say where I want to go and don't touch the controls.. it backs out, drives me, parks and I get out. Every day, thousands and thousands of FSD miles a month on a car I bought in 2023.
>Ten years later not only is it still a broken promise, they're actively working to ship a product that isn't what they promised.
To the contrary, again, what he actually said was:
"As the technology matures, all Tesla vehicles will have the hardware necessary to be fully self-driving with fail-operational capability, meaning that any given system in the car could break and your car will still drive itself safely. When true self-driving is approved by regulators, it will mean that you will be able to summon your Tesla from pretty much anywhere. Once it picks you up, you will be able to sleep, read or do anything else en route to your destination. You will also be able to add your car to the Tesla shared fleet just by tapping a button on the Tesla phone app and have it generate income for you while you’re at work or on vacation, significantly offsetting and at times potentially exceeding the monthly loan or lease cost. In cities where demand exceeds the supply of customer-owned cars, Tesla will operate its own fleet…"
These are all factual statements.. given we've already established aggressive timelines that don't match reality, it really is just beating a dead horse.
Deriving statements like
"The basic news is that all Tesla vehicles exiting the factory have the hardware necessary for Level 5 autonomy… Every car we make… are shipping now with Level 5, meaning hardware capable of full self-driving, or driverless, capability. "
and
"Our goal is, and I feel pretty good about this goal, that we’ll be able to do a demonstration drive of full autonomy all the way from LA to New York, from home in LA to let’s say dropping you off in Times Square in New York, and then having the car go park itself, by the end of next year. Without the need for a single touch, including the charger.”"
The Banish (drop you off and park) is not publicly available yet (regulations/operational) and wireless charging is just now being rolled out.. but again, the substance of his statement is true and will evolve over time to be fully realized.
So if you're arguing he has "pathological optimism" of overstating delivery times, I agree fully.
But suggesting he is lying or intentionally breaking promises within his control, I don't agree.
>This isn't a "lie" to plan for something and have those plans fall through, the intent was to deliver and it didn't happen.
It is if you say it and people buy your product and do not get what they paid for. Everything Elon has said have to be delivered to the buyer. Otherwise it is a lie - false marketing - which is also illegal. You should be paid by him with the time you use at protecting him from his own crap.
>It is if you say it and people buy your product and do not get what they paid for.
If you read through my reply and the specific statements that Elon or Tesla has made, with the exception of HW2.5 and early HW3 (in which they provided monetary compensation in terms of enhanced trade in and upgrade options), statements beyond that are pretty clear to be future and forward thinking. I specifically quoted them above.
>You should be paid by him with the time you use at protecting him from his own crap.
I'm simply articulating the facts of what have actually been said and done without the emotional interpretation. Terms like "lie" and "false marketing" and "illegal" have specific meaning.
I don't want to rehash all of the actual quotes they've made, I did that in a prior reply. Those statements stand on their own and I'm not defending them in any way, only to say that with exception of the 2.5/early 3HW, they did not read as "promises or guarantees" or even definitive timelines in any way.
There is an active case to determine "claims accusing the electric vehicle maker of misleading them about the capabilities of its full self-driving technology", that as of this morning the US appeals court (a three-judge panel of the San Francisco-based 9th US Circuit Court) did not qualify for class action status. So legally the courts would be the right venue to make that determination.
"CLAUDE.md files solve this by giving Claude persistent context about your project. A well-configured CLAUDE.md transforms how Claude works with your specific project. The file serves multiple purposes: providing architectural context, establishing workflows, and connecting Claude to your development tools. Each addition should solve a real problem you have encountered, not theoretical concerns about what Claude might need."
I pretty clearly say I still use CLAUDE.md/AGENTS.md
I've just been trimming down to the recommended < 100 lines and I keep only super high level info in there and allow the agents to figure things out by reading the code.
I wouldn't use a blog post from Nov 2025 be my guiding light.
Anthropic over the past few months have said they've cut their system prompts 80%. These are current best practices. The models are good enough and can figure things out on their own BETTER than you can present the info in markdown files and "decision" documents.
biased opinions andd disucssion are fine, in this instance, a link to the same story as a primary source with a non editorialized title is what I was suggestion from Dang or whoever it is that makes those changes these days.
Guess I'll take my negative 4 votes for suggesting that reuters may be a better primary source. ¯\_(ツ)_/¯
There are a variety of other ways that he could schedule triggered sales that aren't immediately transparent to the media.
A lack of 10b5-1 does not directly correlate to not having any pre scheduled plan.
The fact that they're acquitting months to weeks, makes me feel like they're inventing news to throw on the pile of Nvidia headlines.
There are defined "open windows" he can sell, and these were sold in those permitted window.
It was also multiple batches over months.
He gave form 144 notice and portions were through trusts and charitable / foundation.
He filed multiple Form 144s notifying the SEC of proposed sales under Rule 144. Notable examples include a September 2026 notice for up to 5 million shares (roughly $1.09 billion) and earlier notices (e.g., involving 1.7 million shares in June 2026).
He's got this exact same pattern of behavior for the last 43 sales dating back to 2021.. it's the furthest that I left before I stopped.
reply