This “mandatory meeting” is just the usual weekly company-wide meeting where recent operational issues are discussed. There was a big operational issue last week, so of course this week will have more attendance and discussion.
This meeting happens literally every week, and has for years. Feels like the media is making a mountain out of a mole hill here.
>He asked staff to attend the meeting, which is normally optional.
Is that false? It also discusses a new policy:
>Junior and mid-level engineers will now require more senior engineers to sign off any AI-assisted changes, Treadwell added.
Is that inaccurate? It is good context that this is a regularly scheduled meeting. But, regularly scheduled meetings can have newsworthy things happen at them.
When an SVP asks you to do something in a mass email, it's very much optional. Dave Treadwell is an SVP, his org is likely in the 10's of thousands, there is no way to even have a mandatory meeting for that many people.
My SVP asks me to do things all the time, indirectly. I do probably 5% of them.
> org is likely in the 10's of thousands, there is no way to even have a mandatory meeting for that many people.
Ok, this is pretty off-topic, but is this still true? I get that you can't have 10K people all actively participate in the meeting at the same time, but doesn't Zoom have a feature where you can broadcast to thousands and thousands?
Doesn't X/Twitter have a feature like this? (Although, to be fair, the last time I heard about that it was part of a headline like "DeSantis announcement of Presidential run on X/Twitter delayed for hours as X/Twitter's tech stack collapses under 200K viewers")
But still - nowadays it seems like it should be possible to have 10K employees all tune in at the same time and then call it a meeting, yes?
Yes, but at that point it's an all-hands presentation, and you are basically doing a very careful presentation, thinking about every minute, because of how many hours the "meeting" is costing you.
Very different from the typical weekly/montly outage meeting, where discussion is actually expected, instead of being a ritual.
> but doesn't Zoom have a feature where you can broadcast to thousands and thousands?
They have webinar/event support for 5000+ participants, viewers can raise hands/use chat feedback for questions etc. and the meeting host can invite people to be visible.
The meeting isn't the hard part—after all, shareholder meetings have huge audiences too. Enforcing mandatory attendance for myriads of employees is the hard part, so it's more likely mandatory in name only.
Right, but if you say something essential in a meeting with 10 people and it has to percolate through five levels of management to reach the front-lines and gets watered down, that could be much more lost, even millions.
Scale cuts both ways.
What matters isn't how big the meeting is, it's how important the material is, and how well presented it is.
I don't think I've ever heard a top leader say anything essential in such a meeting. The stuff they work on is not related to my job at all. It's all gartner level strategy stuff. In our company they do take time talking about it in large calls but it's always boring and never relevant. And a lot of political spin you have to poke through to see the real message.
If I ever attend it just put it on mute and look at the slides while I do some real work. That way my attendance gets registered and it doesn't stress me out later with too much stuff left hanging.
That percolation is also translation of what they say to things that are relevant at my level. Like what we will be working on next year, if there's going to be bonus or job losses.
I couldn't give a crap about the company's strategy as a whole and that's not my job anyway. Why should I. I'm not here because I believe in some holy mission. I just wanna do something I like and get paid.
Most of those meetings are pretty damn fluffy. No one goes back to their desk and does anything different because they've introduced new company values and the acronym is S.M.I.L.E.
But this meeting is a course correction for how they're using AI, which is a huge initiative. He'll be trying to sell the right balance of "keep using the technology, but don't fuck anything up."
Too cautious, everyone freezes and there's a slowdown[0]. Too soft, everyone thinks it's "another empty warning not to fuck up" and they go right back to fucking everything up because the real message was "don't you dare slow down." After the talk, people will have conversations about "what did they really mean?"
[0] If you hate AI, feel free to flip the direction of the effect.
Well this is the main problem with AI right now isn't it? How to use it successfully without having it fuck up.
How are they expecting some juniors to do this when the industry as a whole doesn't know where to begin yet?
Like that Meta AI expert who wiped her whole mailbox with openclaw. These are the people who should come up with the answers.
Ps I mostly hate AI but I do see some potential. Right now it feels like we're entering a fireworks bunker looking for a pot of gold and having only a box of matches for illumination.
What we need to know from management is exactly what you mention. Do we go all out and accept that shit will hit the fan once in a while (the old move fast and break things) or do we micromanage and basically work manually like old. And that they accept the risk either way. That kind of strategy is really business leader kind of work. Blaming it on your techs when it inevitably goes wrong is not.
Because the tech as it is right now is very non-deterministic. One day it works magic and the next day it blows up.
And yes that SMILE thing was a good example. Been in too many of those time wasters.
Unless that 30-second stupid joke is what gets the audience to take your request seriously. Sometimes people will help you when you don't come across like a self-interested corporate tool.
I have never in my long life heard a joke from upper management during a meeting/presentation that wasn't awkward and cringe. Just get to the point - tell us how many people are getting fired, so the people who aren't fired can get back to work, and you go back to running this company into the ground.
I was thinking about this in recent weeks and I think I’ve actually changed my mind on it.
It’s not really possible to measure how much it would cost to not have a meeting, and I think it’s pretty obvious that if there were no meetings ever, it would hurt a company a lot
Yeah, I agree it's a silly metric. But it's kinda also a good reminder that meetings do have a cost associated with them, so they should stay short, focused, and held only when necessary.
"This could have been an e-mail" should never need to be said.
are you saying SVP’s words are not important and should be ignored? This is not what I remember back in the day when Bezos sent his email with a question mark (or maybe !)
> are you saying SVP’s words are not important and should be ignored?
Personally I would say that an SVPs words are not important and don't need to be ignored.
It's like a politician talking about abstract policies. Yes they do sort of affect me, but they don't require any affirmative action on my behalf any more than the wind does.
That's not really what the headline attempts to communicate though. It specifically emphasizes "Mandatory" and "AI breaking things". Nobody was going to click on "Regularly scheduled Amazon staff meeting will include discussion on operational improvement"
But it gets less mandatory the more layers up you go. If I get an email from an SVP that is CC: the entire division saying everyone should go to a meeting I will almost certainly be able to ascertain the contents of that meeting in 10 seconds from someone else who did attend
If it's actually really mandatory, my manager will probably also relay that directly to me. And that resets the count for "less mandatory the more layers up you go".
The bosses function is to shield you from random time wasting junk. Either you haven't had to survive in a borg corporate environment or you have and you had a bad boss for it.
Days are not far, where my agents are going to attend meetings & share my opinions, collect summary for me. If everyone do same - agents run meetings & share summary with parent (humans). Each of us have LLMs/Agents with our contextual data. It is another level of multi tasking.
Then I spin up another agent to listen to the agent who went to the meeting and make any necessary adjustments to the output of my coding agents based on the new rules it heard about from the meeting agent.
My agent will just be full AGI. It’ll invent time travel and go back to attend all my meetings 100x faster.
Meanwhile the normie “Claw/OpenBot” agents can stay in the present grinding 24/7, while mine recursively spawns across alternate timelines and handles my work at ~1e9x parallelism.
Note that the article doesn’t say that he told staff they have to attend the meeting. It says he “asked” staff to attend the meeting. Which again, it’s really really normal for there to be an encouragement of “hey, since we just had an operational event, it would be good to prioritize attending this meeting where we discuss how to avoid operational events”.
As for the second quote: senior engineers have always been required to sign off on changes from junior engineers. There’s nothing new there. And there is nothing specific to AI that was announced.
This entire meeting and message is basically just saying “hey we’ve been getting a little sloppy at following our operational best practices, this is a reminder to be less sloppy”. It’s a massive nothingburger.
That really isn’t the culture at Amazon. There are all-team meetings that happen all the time, and every now and then there is a reminder that “hey we’re gonna be talking about an interesting topic so you might want to join”, but it is certainly not a mandate or expectation that everyone will join.
Different companies have different cultures. Weird that people can’t grok this.
Your characterization of the event as a simple reminder to follow established best practices is directly contradicted by the briefing note of the meeting, which specifically mentions a lack of best practices related to AI. Which makes me skeptical of your assessment of the situation in general.
> Under “contributing factors” the note included “novel GenAI usage for which best practices and safeguards are not yet fully established”.
It’s part of the change management process that all code is reviewed. This is needed as per several different compliance agreements. What’s probably happened is poor peer reviews from other junior engineers gets missed. That’s a lot of code reviews to send upstream.
It didn't seem to make the news but at least in NYC the entire Amazon storefront was broken all afternoon on Friday.
Items weren't displaying prices and it was impossible to add anything to your cart. It lasted from about 2pm to 5pm.
It's especially strange because if a computer glitch brought down a large retail competitor like Walmart I probably would have seen something even though their sales volume is lower.
A little birdie told me someone pushed duplicate data into one of Amazon’s core noSQL systems that runs most of e-commerce. The front end of the site broke in weird ways but it certainly wasn’t taking orders.
Funny, I was automatically refunded for a pair of shoes that Amazon thought I never received even though I’m wearing them right now. I couldn’t even find a way to dispute the refund so I just took the win…
I am not in that specific meeting but it made me chuckle that a weekly ops meeting will somehow get media attention. It's been an Amazon thing forever. Wait until the public learns about CoEs!
Exactly. This is real world pushback on the "software is solved" narrative from AI labs. Also, most orgs try to copy Amazon for some reason more than big tech firms. "At our org, we disagree and commit" - yeah you made that one up yourself. Anyway, this is going to have a lot of impact in my view.
There was nothing mentioned in the meeting or messaging about PRs with AI contributions. There are no extra requirements for review or scrutiny of AI-generated-code. The media reports about this have been excessively misleading about this.
There's an explicit tension: SWEs would love that as a "get out of jail free" card, but their management chain is being evaluated by ajassy on AI/ML adoption. Admitting AI code as the root cause of a CoE is gonna look really bad unless/until your peers are also copping to it.
I think its a question 2 or 3 in a why chain, but 4 and 5 need to be why the agent screwed up, and there needs to be action items that are around giving the ai better guardrails, context, or tooling.
"get a person to look at it" is a cop-out action item, and best intentions only. nothing that you could actually apply to make development better across the whole company
> This meeting happens literally every week, and has for years. Feels like the media is making a mountain out of a mole hill here.
Are you completely missing the point of the submission? It's not about "Amazon has a mandatory weekly meeting" but about the contents of that specific meeting, about AI-assisted tooling leading to "trends of incidents", having a "large blast radius" and "best practices and safeguards are not yet fully established".
No one cares how often the meeting in general is held, or if it's mandatory or not.
>> Are you completely missing the point of the submission
no, and that's what people are noting: the headline deliberately tries to blow this up into a big deal. When did you last see the HN post about Amazon's mandatory meeting to discuss a human-caused outage, or a post mortem? It's not because they don't happen...
Amazon has had a really bad string of various outages recently. Assuming they're internally treating this as business as usual in post-mortems then perhaps the newsworthy thing is actually that they aren't taking their outages seriously enough.
> the headline deliberately tries to blow this up into a big deal
I do not understand how “company that runs half the internet has had major recent outages and now explicitly names lax/non-existent LLM usage guidelines as a major reason” can possibly not be a big deal in the midst of an industry-wide hype wave over how the world’s biggest companies now run agent teams shipping 150 pull requests an hour.
The chain of events is “AWS has been having a pretty awful time as far as outages go”, and now “result of an operational meeting is that the company will cut down on the use of autonomous AI.” You don’t need CoT-level reasoning to come to the natural conclusion here.
If we could, as a species, collectively, stop measuring the relevance of a piece of news proportionally by how much we like hearing it, please?
And too many people have their egos tied to its failure, too.
Im a massive AI skeptic. If anyone were to be jumping up and down on the corpse of AI and this incessant drive to use it everywhere, it’d be me. But I also work at Amazon. I got the email. I attended the meeting. I can personally attest that there are no new requirements for AI-generated code. The articles about this in the meeting at extremely misleading, if not outright wrong. But instead of believing the person that was actually there in the room, this thread is full of people dismissing my first-hand account of the situation because it doesn’t align with the “haha AI failed” viewpoint.
Not just their egos, but their paychecks. This place is either going to get very quiet or really weird when the hype train derails and the AI bubble bursts.
The subject of the media coverage is not AWS, it is a peer organization to AWS that runs using significant amounts of non-AWS infrastructure. They are both part of an umbrella called Amazon but are not at all the same thing.
It's hard to that this objection seriously. The publication is literally called the Financial Times. It's not exactly crazy for them to think that their readers might care about the entity that shows up the stock ticker rather than how the company happens to divide up things internally.
Even if it weren't a finance publication, I have trouble imagining you making this argument if a headline said something like "Google deals with outages in the cloud" because of the idea that it's misleading to refer to it as anything other than GCP. I think you're fundamentally not understanding how people communicate about this sort of thing if you actually think that someone saying "Amazon" is misleading in any meaningful way.
You’re describing reasonable misunderstandings, but they are still misunderstandings.
The cause and effect statements just don’t correspond to reality.
I guess I’m stuck on the idea that the actual facts are relevant. If the question instead is how the dance of optics and PR is going in the minds of people who don’t know enough to doubt what they read, I don’t know what to say about that.
The message and meeting being discussed here have nothing to do with AWS or any outages AWS has faced recently. I think you’re missing the point of the discussion.
I don’t blame you, because this is just bad reporting (and potentially intentionally malicious to make you think it’s about AWS). But the meeting and discussion was with the Amazon retail teams, talking about Amazon retail processes, and Amazon retail services. The teams and processes that handle this are entirely separate from any AWS outages you are thinking of.
The outages that Amazon retail has faced also have nothing to do with AI, and there was no “explicit call out” about AI causing anything.
The core message of the article is that Amazon has been having issues with AI slop causing operational reliability concerns, and that seems to be 100% accurate.
I believe it is by group - AWS started the weekly operations meeting, effectively every service's oncall from the last week had to attend. Then it grew massive, so they made it optional. Alexa had a similar meeting that tried to replicate what AWS did. A lot of time spent reviewing load tests getting ready for holiday season, prime day, and the superbowl (super bowl ads used to cause crazy TPS spikes for Alexa). And a lot of finger pointing if there was an outage from one team. While it probably did help raise the operational bar, so much time wasted by engineers on busywork/paperwork documenting an error or fix vs improving the actual service.
None of the hyper scalers are going to stop offering Claude. All of the big 3 have invested billions of dollars into Anthropic, and have tens (if not hundreds) of billions more tied up in funding deals with them. Amazon and Google are two of the largest shareholders of Anthropic.
Anthropic is going to be fine. The DoD is going to walk this back and pretend it never happened to save face.
Tens, maybe hundreds, of billions? That’s cute. The DoD will spend $961b this year. It does that like clockwork every year, year after year.
Anthropic is not even close to too big to fail. And even if this could get settled in court 5 years from now, this can easily throw enough of a wrench into their revenue streams to kill their flywheel.
This isn’t “a few billion”. Maybe you missed some of the earlier comments. The hyperscalers have hundreds of billions of dollars tied up in deals with Anthropic. You’re delusional if you think these boards aren’t going to have a back room talk with Hegseth to smack some sense into him. This gets walked back next week, guaranteed.
The counterparty risk on those buildout contracts is not the same as their equity investments. Amazon isn’t assuming the entirety of that buildout exposure as a vote of confidence or form of investment in anthropic; they’re hedging it with insurance, credit default swaps, and MAE clauses.
Those datacenters are AWS infrastructure that Amazon owns and can repurpose. The equity stake is the only part that’s truly at risk, and $8B is a rounding error on Amazon’s balance sheet.
That $961 billion includes things like airplanes and bullets, tech companies are only getting a taste of that pie not anywhere close to the whole thing.
Ironically, of all things Trump has done so far, closing Anthropic could set a new record for pissing off the highest number of people globally. Outside of HN with a group of dedicated people who is against it, the whole global software world is already running on CC.
Half of that budget gets contracted out to Lockheed, Raytheon, Northrop, Boeing, General Dynamics, etc. Those companies absolutely do spend money on the hyperscalers.
obviously, I was never suggesting that the DoD spends $961b a year on cloud computing.
Look, it’s a very simple question: Amazon has invested $8b into anthropic. Do you think if the DoD disappeared tomorrow that Amazon would lose more than $8b in revenue over the next 5 years?
I think you underestimate how large the DoD budget is and how many times that money changes hands in the pursuit of fulfilling contracts. $20b-$25b in revenue per year across all hyperscalers is a totally reasonable estimate.
I would find that a lot more plausible if people had not spent the past week giving me similar arguments, in precisely the same tone, for why this was an empty threat and would never happen in the first place. If Amazon and Google do not either bow down or immediately join a business coalition to get Trump out of power, Hegseth will be even happier to get an opportunity to prove his power by destroying them. Trump either doesn't want to stop him or has become too senile to stop him.
GovCloud revenue is in the tens of billions of dollars. Bedrock less so. Almost every FedRAMP product uses the same codebase for Fed and non-Fed, and this would force most FedRAMP vendors to blackball Anthropic.
This restriction is viral. If AWS hosts Claude models, Lockheed can no longer use AWS for anything. Every defense contractor will pull out. What if Lockheed uses Asana or Jira or Slack? Guess what, they better not use Claude ANYWHERE in their organizations, or else all defense contractors will have to drop these products. Any any other company whose product they use in the design or manufacture of their products - if anyone, anywhere is using Claude products, they have to be dropped.
The JWCC, which is larger than GovCloud, was only $9b, split across three companies, over ten years. It’s peanuts compared to the investments that the hyperscalers have with Anthropic.
JWCC is not the only project. Vendors like Crowdstrike also rely on hyperscalers to serve their products to federal customers, and the codebase is shared.
This announcement has made Anthropic toxic in the entire dependency chain because it means years of efforts and tens to hundreds of millions of dollars rearchitecting entire platforms and renegotiating contracts.
The entire cybersecurity industry has a TAM of $208 BILLION [0]
> because it means years of efforts and tens to hundreds of millions of dollars rearchitecting entire platforms and renegotiating contracts.
This is exactly why this announcement has not made Anthropic toxic. The entire industry knows how ridiculous this move is from Hegseth, and it’s going to be rolled back next week once the adults get back from their weekend.
It will really depend on the fine details. If Amazon would lose its military contracts unless it dropped Claude, then Claude will be gone tomorrow. They just got a half billion contract for the Air Force earlier this year, and it's not their only military contract, and they're going to want to be well positioned next time something like the JEDI contract comes along.
Also, AWS has a long history of rolling over when politicians make noise about AWS customers, going back to when Joe Lieberman casually asked Bezos to please stop supporting Wikileaks.
I don't think you understand. This supply chain risk designation is viral. Every Claude model provider now has to decide whether to (1) drop Anthropic models, or (2) drop every single government contract, every contract with government contractors, or any customer who has any customer to any degree of connection to a government contract [which is effectively everyone], or (3) go to jail.
AWS/GCP/Azure all do business with the DoD and at least AWS and Azure use Claude a decent amount internally. AWS’s Kiro tool (which is used internally instead of Claude Code) relies entirely on Claude models.
This is almost certainly going to be rolled back, because I guarantee the DoD isn’t going to stop doing business with the hyper scalers, and the hyper scalers aren’t going to stop doing business with Anthropic.
In the before-AI world, it mattered a lot where data centers were geographically located. They needed to be in the same general location as population centers for latency reasons, and they needed to be in an area that was near major fiber hubs (with multiple connections and providers) for connectivity and failover. They also needed cheap power. This means there’s only a few ideal locations in the US: places like Virginia, Oregon, Ohio, Dallas, Kansas City, Denver, SF are all big fiber hubs. Oregon for example also has cheap power and water.
Then you have the compounding effect where as you expand your data centers, you want them near your already existing data centers for inter-DC latency reasons. AWS can’t expand us-east-1 capacity by building a data center in Oklahoma because it breaks things like inter-DC replication.
Enter LLMs: massive need for expanded compute capacity, but latency and failover connectivity doesn’t really matter (the extra latency from sending a prompt to compute far away is dwarfed by the inference time, and latency for training matters even less). This opens up the new possibility for data centers to be placed in geographic places they couldn’t be before, and now the big priority’s just open land, cheap power, and water.
>Oregon for example also has cheap power and water.
Cheap for who? For the companies having billions upon billions of dollars shoved into their pockets while still managing to lose all that money?
Power won't be cheap after the datacenters move in. Then the price of power goes up for everyone, including the residents who lived there before the datacenter was built. The "AI" companies won't care, they'll just do another round of funding.
I travel for work often (used to do it every week but now it’s once-ish a month), and fly business every now and then. I don’t think I’ve ever met any fellow work flyers who wanted the flying experience to be more focused on the “business aspect”. The lounge is for relaxing, and the comfortable seat on the plane is so I can sleep and not be a zombie when I land. I’ll work when I get to the destination, not while traveling.
And the original link about investment in India is also about fulfillment jobs and even worse, “investing in AI”, aka building data centers, which contribute essentially no jobs at all.
Where are you seeing “American” jobs? Amazon workers in India were laid off too.
There are similar stories about Amazon investing in American cities too. Cherry picking a story that Amazon is renovating their office in India is ingenuine.
At minimum this thing should be installed in its own VM. I shudder to think of people running this on their personal machine…
I’ve been toying around with it and the only credentials I’m giving it are specifically scoped down and/or are new user accounts created specifically for this thing to use. I don’t trust this thing at all with my own personal GitHub credentials or anything that’s even remotely touching my credit cards.
That list of movies is just the movies that Amazon Studios has been the distributor for via Prime Video. Amazon didn’t necessarily produce or fund all of the movies in that list. It’s a bunch of cheap movies that are likely meant to be loss leaders for Prime Video subscriptions, which is something that very much does fit the style of Amazon. Netflix, Hulu, Apple TV all have a similar list of D-tier garbage just to fill their catalogs.
On the contrary to your points, Amazon has put out some pretty solid and well received original series. The Boys, Gen V, Fallout, Reacher, Mr and Mrs Smith, Invincible, have all done really well if not been hits.
Games is pretty trash though. I think they’re also going for a loss leader strategy there, but the platform they’re trying to promote (Luna) just isn’t there.
> it uses the Claude models but afaik it is constantly changing which one is using depending on the perceived difficulty
Claude Code does the same. You can disable it in Kiro by specifically setting the model to what you want rather than “auto” using /model.
Tbh I’ve found Kiro to be much better than Claude Code. The actual quality of results seems about the same, but I’ve had multiple instances where Claude Code get stuck because of errors making tool calls whereas Kiro just works. Personally I also just prefer the simplicity of Kiro’s UX over CC’s relative “flashy” TUI.
This meeting happens literally every week, and has for years. Feels like the media is making a mountain out of a mole hill here.
reply