Imagine you walk into a bakery, and you ask for a dozen donuts, only to receive 11 donuts. You complain that you asked for 12 donuts and only received 11, but the baker looks at you and says “No, lad. These days a dozen is only 11. You asked for a dozen and I gave you a dozen”. Or you go to a car salesman, and he tells you about a 275 horsepower vehicle he has on offer - that sounds fast! - but when you give it a test drive, it putters along at a snail’s pace. You’d be confused, right? But then he looks you in the eye, and says “Well, but horsepower doesn’t mean what it meant back in the day. Haven’t you been keeping up with the standards published by the government?”
In theory, we could define units to mean whatever we want, and we could make our units have different values from year to year. But in general, we don’t do that, because doing so is stupid - we use units to help us understand and quantify what’s going on, and changing the definitions of units just makes things confusing and makes it harder to keep track of the world. But the one measure which we still do this with is value: one of the functions of a currency is as a unit of value, but the amount of value represented by a dollar today isn’t the same as it was 50 years ago, or what it will be 50 years from now. Economically literate people are already good at adjusting for inflation, which is a hack to get around this, but people do make mistakes failing or forgetting to adjust for inflation, and an ideal unit of value shouldn’t be changing constantly over time.
Note that I’m not saying a currency shouldn’t be inflationary - there are good reasons for currencies to be inflationary (this is particularly true for UBI coins, but also in general) which I won’t get into here, but inflation can be achieved without tampering with the unit of measurement of value - instead of increasing the supply of the coin, you can make user’s accounts decay, so if someone is holding 100 currency at one point in time, they may hold 97 currency two years later, assuming that the account is left untouched. The side effects of doing this are the same as if inflation is implemented in the usual way, but leaves the unit of value untouched. 
There are two ways to approach the stabilization of the measurement of value: Either you can hold the total supply of currency fixed, or you can try to track a basket of goods, and maintain a constant exchange rate between that basket and the currency. Tracking a basket makes the most sense if your primary goal is to create a perfectly stable unit of value, however this removes an important dynamic in the modern ecosystem of digital currencies: Many digital currencies become popular due to people investing in the currency in the hope that the value of the currency will rise over time. It’s worth noting that even an inflationary currency can be a good investment, if the rate of inflation / decay is less than the rate of adoption of the currency. As an example, 1 bitcoin was worth $414 five years ago, in March 2016, but is worth $51,000 today, for a annual increase in value of 160% per year. Even if bitcoin inflated at the same rate as the USD (ignoring second-order effects of inflation on price, which do actually matter in the case of BTC), this would have been an amazing investment.
Ideally, you want the unit of value to be stable in the long run, but also to allow flexibility for the currency to grow in the short run, which is why it makes sense for even inflationary currencies to aim to have fixed total supply of currency.
 It makes sense that this isn’t usually done with paper currencies, since it is cumbersome to keep track of decay with paper (although it has been done using costed stamps that are required to maintain the validity of paper bills); but it’s much easier to decay the value of an account with a digital currency
The most important reason for an inflationary currency is to defeat sticky prices and money illusion. Destroying the value of savings over time is not the goal.
I don't think the typical person arguing for inflation would agree that "Destroying the value of savings over time is not the goal"- this is the reason that I have most commonly heard cited for inflation, since it ensures that people spend their money and engage in trade.
Having done some more reading, I agree that overcoming the cognitive bias where people don't want prices to change is an important use of inflation, and I agree that my proposal does not succeed in addressing this, since the entire point there is that by changing the value of a currency, you can get people to accept lower de facto wages without lowering their nominal (perceived) wage. I will point out that this currently only works in one direction, and does not help people to accept paying higher prices for goods (aside from more quickly forcing sellers to the point where they have to raise prices, at which point they may just as well set a reasonable de facto price that consumers would not otherwise have accepted). I also note that this usage feels somewhat dishonest to me, since it inherently works by creating a mismatch between people's intuitive maps of the world and the actual reality of the situation.
That's only partly the case. Stable coins exist and there are currently four with a market cap higher then a billion dollars. Most of the stable coins are currently pegged to the US dollar but there's no reason why you couldn't have a stable coin that's pegged to whatever basked of goods you care about.
If you want to make bets in a prediction market denoting them in a currency that doesn't inflate might be worthwhile.
I would be surprised if we don't have stable coins pecked to a goods basket in a few years when the DeFi field matures.
I very much agree with this
A fixed supply doesn't guarantee a stable price. Depending on people's need for the currency it can still fluctuate in price.
Wouldn't a currency whose total supply is fixed while the economy grows, lead to the same problems with changing definition of horsepower and so on?
Correct- as I stated in my post, if it's desired to have zero fluctuation in value, the currency should be pegged to a basket of goods, which implies a variable total supply; however in cases where some amount of variance in value (generally less than what currently exists with USD) is acceptable, then a fixed total supply can make sense
That reminds me how gigabyte shrunk from 1073741824 bytes to 1000000000 bytes, so you would buy a "500 GB" disk, put it into your computer, where the operating system says it's actually 466 GB, but whatever.
I don't think this is an example of definitions changing over time, but just where the same word has always meant two different things in different (closely related) uses. My understanding is that marketers for manufacturers always used the decimal meaning to market drives, while everybody else uses the (slightly larger) binary value
One of the cornerstones of Conservative thought is the following argument:
The United States (and the Western world in general) have consistently been the most productive and best nations, out of all the nations on earth, since at least the 1940s. Ergo, the best policy is to revert to what is known to work.
This has obvious arguments for correctness.
One of those things they all have is an inflationary money supply. There are no working examples otherwise. The thing that "Progressives" in the United States is they want to copy elements of working example countries, notably Nordic countries, for things like healthcare/criminal justice/taxation/social welfare, etc. Progressives aren't asking to change everything, they just want to import what the evidence says are superior methods in specific areas of governance.
Do you know of a working example where this has been tried? Because messing with the money supply has enormous risks, we already know the economic system has flaws that will cause it to periodically collapse.
After reading your response, I notice that I made a mistake in my original post in terms of communicating my thoughts. This was originally written as part of a larger piece about the design of new currencies, in the gestalt of the current thriving cryptocurrency scene, and in that context it was clear that I was not advocating for reforming currently existing currencies along these lines, but rather presenting thoughts that designers of future, non-state-backed digital currencies should take into consideration - a space where experimentation without existing working examples is well-justified - however that context is not provided in this post, which can make it look like I am advocating reforms more extreme than I actually endorse.
I don't think currency is a -unit- of value, in fact I don't think such a thing can exist. Value for me means, "what would you be willing to give up to obtain something?" Value is subjective; a glass of water generally has close to zero value but if you're stranded in the desert it will have an extremely high value.
And this way value can be created: I have lots of glasses of water in this desert, you have none. You give me your kingdom and I give you a glass and we both give up less value than we get.
This analysis neglects the concept of velocity. If there is a fixed amount of dollars but (for example) a collapse in the housing market causes velocity to plummet, then unexpected deflation will wreck more chaos on the economy than an inflation rate of 1-2%.
A currency with a fixed supply is not stable.