Work Amplification: Difference between revisions
Added page description via AutoDescriptor bot |
No edit summary |
||
| Line 14: | Line 14: | ||
== Conversation == | == Conversation == | ||
A similar thing happens in conversations as well. However, unlike in the client-server model, conversation partners are rarely paying each other for services. Consequently, the optimal strategy in any case is for the requester to abandon the conversation. This is especially true online, where | A similar thing happens in conversations as well. However, unlike in the client-server model, conversation partners are rarely paying each other for services. Consequently, the optimal strategy in any case is for the requester to abandon the conversation. This is especially true online, where it is cheap to ask "source?" and expensive to actually provide a source that is both accessible, reputable, and which matches the political proclivities of every participant. Online commenters frequently espouse the idea that "the burden of proof is on the one making the claim", but realistically they're just asking for education for free. | ||
The website [https://lmgtfy.app/ Let Me Google That For You] was invented for this purpose, but as always [[wikipedia:Eternal September|Eternal September]] never ends. Sometimes people try to help others help themselves but this rarely works. | The website [https://lmgtfy.app/ Let Me Google That For You] was invented for this purpose, but as always [[wikipedia:Eternal September|Eternal September]] never ends. Sometimes people try to help others help themselves but this rarely works. | ||
Revision as of 07:36, 15 November 2025
Work Amplification occurs when a request takes a much smaller effort to make than a response. This is a problem in interactions both in the software client/server and in a human context. Any low-effort ask that requires a high-effort response could be work amplification.
Software
In software, a host of techniques exist to reduce the scope of the problem:
- caching - cut subsequent response costs at the price of freshness
- debouncing - assist with accidental request replication
- rate-limiting - control requests to match response costs
- request aggregation - set `n` request cost to same as 1 request cost
These are all necessary because otherwise it's trivial for someone to intentionally or accidentally result in denial of service. Modern designs usually distribute work-execution from work-determination which allows for many of these things to not be a problem. You may receive a thousand requests, but you only ever do the expensive thing every minute or so which allows you to just aggregate them.
Conversation
A similar thing happens in conversations as well. However, unlike in the client-server model, conversation partners are rarely paying each other for services. Consequently, the optimal strategy in any case is for the requester to abandon the conversation. This is especially true online, where it is cheap to ask "source?" and expensive to actually provide a source that is both accessible, reputable, and which matches the political proclivities of every participant. Online commenters frequently espouse the idea that "the burden of proof is on the one making the claim", but realistically they're just asking for education for free.
The website Let Me Google That For You was invented for this purpose, but as always Eternal September never ends. Sometimes people try to help others help themselves but this rarely works.
"and extra especially nuclear, in addition to making power lines difficult to build, are to blame"
I used to think nuclear reactors are just hard to build in general, because the costs when something goes wrong are very, very high. So what unnecessary regulation is there with nuclear reactors that you think should be deleted?
I'm sorry but this is the easiest thing to google in history, don't make people do the work for you.\n\nStart here:\n\n1. How many new nuclear power plants has the NRC approved in its entire history (since being formed from the AEC)?\n\n2. What's the cost of a nuclear kw in China vs the US, and is the trend going up or going down?
Perhaps we can learn more on how to manage these online conversations by seeing how we fix them in computer systems.
