That is not something that’s necessary for all CQRS systems, but maybe is something you’ve heard for the subset that people call “Event Sourcing”? There it’s a design goal that the system only records events that are occurring, so there’s no domain level validation that can be done on the command path - the user pressed the button whether we like it or not, so to speak. Whether the event has the intended effect is worked out after the event is recorded.
But there’s nothing in the more general idea of “separate reads from writes” that mandates “no validation on writes”
Commands can validate their input in CQS. What they don't do, in strict CQS, is return values. They can set state which can then be queried after execution which can let you retrieve an updated result or check to see if an error occurred during execution or whatever.
>All of our tests and benchmarks account for repeatability.
What does repeatability have to do with intelligence? If I ask a 6 year old "Is 1+1=2" I don't change my estimation of their intelligence the 400th time they answer correctly.
>The machine in question has no problem replicating its results on whatever test
What machine is that? All the LLMs I have tried produce neat results on very narrow topics but fail on consistency and generality. Which seems like something you would want in a general intelligence.
>What does repeatability have to do with intelligence? If I ask a 6 year old "Is 1+1=2" I don't change my estimation of their intelligence the 400th time they answer correctly.
If your 6 year old can only answer correctly a few times out of that 400 and you don't change your estimation of their understanding of arithmetic then, I sure hope you are not a teacher.
>What machine is that? All the LLMs I have tried produce neat results on very narrow topics but fail on consistency and generality. Which seems like something you would want in a general intelligence.
No LLM will score 80% on benchmark x today then 50% on the same 2 days later. That doesn't happen, so the convoluted setup OP had is meaningless. LLMs do not 'fail' on consistency or generality.
Speaking to US laws, auto manufacturers are required to fix design bugs that cause safety issues regardless of warranty or used status, at no cost to the owner. You may be familiar with the standard name for those fixes, "recalls". It's illegal to sell a vehicle with unresolved recalls, though the government deliberately avoids enforcing that as aggressively as they could.
It's a very different system from software's "NO WARRANTY OF ANY KIND".
The American version "Breakfast is the most important meal of the day" started as a breakfast cereal marketing line, not an endorsement for eating an "English breakfast".
I will take the science over old wives tales and marketing bs any day. Even if the science is as poorly done as most nutrition science is.
I always thought the emphasis on breakfast originated in cultures or times where it would be followed by a long day of manual labor.
This also ignores that most European countries emphasize breakfast even less than Americans, and that the British (home of the English breakfast) have a lower life expectancy than most of their western European peers (but not as low as Americans).
"Slower growth: Despite the record output, some analyses reported that the overall growth of U.S. manufacturing output was modest, at just 1% in 2024. This reflected a lag between announced investments and new operational capacity coming online."
While that might be true, the five whys is notorious for slipping into a destructive "you/I suck and firing you/I solves the problem for good and I believe it makes everyone absolutely happy" style of false conclusions.
Reportedly Toyota has organizational mitigations for that problem or reportedly the working culture there isn't so great after all. The bottom line is, it's a double edged sword to say the very least.
A good habit to practice is to see how far you can go reconciling apparent contradictions with charitable interpretation. I think in this case, I can see "brand loyalty" on a continuum ranging from "feels good about product" to "so completely loyal that lock-in would be redundant". The furthest extreme would produce an effective contradiction, but anything short of that can make sense of the term while leaving space to understand lock in as a rational, or at-least non-contradictory action.
I think that can backfire spectacularly, as we're seeing with Synology, but I suspect that a non-trivial amount of the time, it simply happens and works, no revolt is staged, and profits flow (for better or worse).
The example coming to my mind right how is Pitney Bowes, which sells big envelope stamping and sealing machines. They sell a proprietary sealing fluid (wtf) that, as far as I can tell, is water with blue food coloring. And a costly proprietary red ink cartridge for stamping. But people sign the contracts and the world keeps on keeping on.
> Isn't that a contradictory position? Locking in raises the cost of disloyalty, loyal customers (by definition) don't need to be locked in.
In this case, the customers were loyal to Synology for the NAS but not the hard drives.
By locking them in further, they thought they could capture their customers' hard drive purchasing, too. They thought the brand loyalty would allow it.
You see this a lot when a company’s founders leave and are replaced by MBAs. Customer goodwill isn’t a tangible asset, so the MBAs burn it to produce more quarterly revenue. It works great for a while until the customers wisely decide never to let that company burn them again.
By that stage the MBAs have scored even higher paying jobs at bigger companies based on how much they boosted profits, so I guess it continues to work for them afterwards too.
If you (as the SK government) were going to do a deal with " AWS/GCP/Azure" to run systems for the government, wouldn't you do something like the Jones Act? The datacenters must be within the country and staffed by citizens, etc.
Microsoft exec testified that US Govt can get access to the data Azure stores in other countries. I thought this was a wild allegation but apparently is true [0].
Nope, it was pretty straight forward. It was the 1968 National Flood Insurance Act. Flood prone communities get flood insurance in exchange Gov gets to manage building standards to reduce damage caused by flooding. It was pretty similar to what we are seeing today. There was a hurricane and the Mississippi river flooded. Private insurance started pulling out so the Government decided to step up.
reply