HNHacker News
TopNewBestAskShowJobs

mwsherman

3,607 karma · joined July 21, 2009

https://clipperhouse.com/
submissionscomments
mwsherman··on Lazy Redis is better Redis
As I was reading through, the back of my mind was saying “Nice heuristics but now you’re adding uncertainty. The behavior of DEL could vary dramatically and users would not know why.”

I was relieved that antirez recognized this as a semantic change and gave it a new name. Very thoughtful.

Perhaps there is something to be learned from GC algorithms in this case?

mwsherman··on Hit Charade
No mention of the Brill Building [1]? Pop music has always been factory-produced.

Also, multiple hooks per song is a better song. We like pop music for the hooks. If we don’t like the hooks, we don’t like pop music.

[1] https://en.wikipedia.org/wiki/Brill_Building

mwsherman··on Announcing TypeScript 1.6
Type guards offer just a bit of magic, similar to type switches in Go: https://golang.org/doc/effective_go.html#type_switch
mwsherman··on More shell, less egg (2011)
The key bit of why we do things like this: “useful for testing the value of the answers and for smoking out follow-on questions”.

That statement is not about Knuth’s presumably beautiful work. It’s that the least deterministic (and most important) part of programming is discerning the value of doing the thing at all.

mwsherman··on YC Fellowship
Well put. Geography is vastly underestimated as a limiting factor. Not simply for $ – living in the Bay Area is barely feasible for a young person without dependents – but also the many quality-of-life trade-offs that come with a move. I hope this is a successful experiment.
mwsherman··on What’s the best way to spend $20K to help the common good?
Yup. Now, logistically, how hard is this? It seems promising.

Let’s assume to do this we’d want to dedicate at least one person full-time to this. There’s the sourcing of the iodized salt, determining how one ships such a thing (presumably overseas), and to what specific place it should go. I assume there’s customs to get the shipment across. Then, once it’s there, how it should be distributed.

Phew, that’s sounds like more than one person’s worth of work, and that’s just the first shipment. We need to repeat the process – including raising more funds.

I assume this person that’s hired is desirable for their logistical skills. Let’s imagine they already have money in the bank, and so will work for less than the market would otherwise pay them. How about $60K/year? Add supplies, taxes and benefits to that, I can see $100K per year for this rare value of an employee.

How far did our $20K go? Not very. Everyone is skilled and honest and working hard, but we find that $20K doesn’t supply $20K of iodized salt to people in need.

In fact, it mostly went to the salary of the skilled Westerner who is our organization. Now we see why people create charities, but also why it’s hard for charities to meet their mission even when things like malaria netting are cheap.

To be clear, this is not cynicism. It’s just that if it were a simply problem of multiplying dollars times cheap-but-effective product, it would already be solved.

mwsherman··on Abstractions my Deep Learning word2vec model made
If I understand your question, the idea is that one can do “arithmetic” on concepts. Essentially the first equation asks “Obama : USA :: ? : Russia”. Similarly, “King : Man :: ? : Woman”.

The way the corpus “talks about” Obama in relation to the USA is similar to how the corpus talks about Putin in relation to Russia. That the system can reveal this is amazing to me.

mwsherman··on Yagni
Yes! And in this case, you “cursed” yourself for not making the prediction that this change would be necessary. One discounts the 100 other wrong predictions that would have accompanied this right one.
mwsherman··on Ad Blocks’ Doomsday Scenarios
I am reminded of the panic over DVRs and fast-forwarding past ads.
mwsherman··on Race conditions on Facebook, DigitalOcean and others (fixed)
It’s impressive that they are able to fix them so quickly – one needs to imagine they get a non-trivial number of reports, and that some majority of them are junk. They have a good triage + repro + escalation system.
mwsherman··on FBI overstated forensic hair matches in nearly all trials before 2000
Legal arguments, like most things in life, reward narrative over truth. “Truth” is an unbelievably high bar to pass – even basic ideas like confidence level are beyond the understanding of the legal system.

“Beyond a reasonable doubt” is actually a decent epistemic standard. The problem is that juries and judges are not exposed to the information that would lead to reasonable doubt in these cases.

So basic epistemology already makes bad outcomes likely. Add the incentives (political, organizational) in cases like these, and terrible things happen.

mwsherman··on How Heartbleed could've been found
Oh no doubt, but tools scale well. The tools allow us to partially but cheaply replicate the brains of the expensive smart people.
mwsherman··on How Heartbleed could've been found
The third part of my post here [1] is that progress in security will more likely come from cheap tools than expensive humans. Pardon the self-reference of course. :)

[1] http://clipperhouse.com/2015/04/04/liquidity-open-source-and...

mwsherman··on China's Man-On-the-Side Attack on GitHub
Can browsers or OSs not treat the corrupted Baidu analytics as malware?
mwsherman··on Choose Boring Technology
FWIW, maybe the approach in this case is traditional normalization to get the product out the door. When it becomes untenable, hire someone for whom star schemas are boring.

Put another way, the star schema was known to be a better approach only after the fact. Had your team researched “exotic” (unfamiliar) approaches early on, there is no guarantee you would have landed on a star schema.

mwsherman··on How I doubled my Internet speed with OpenWRT
I don't see any benchmarks, did I miss them? I suspect this doesn't double bandwidth, the packets all end up over the same neighborhood network.

The only way it would increase bandwidth is if it evades artificial throttles. Having two connections to the same (neighborhood) pipe seems useless.

mwsherman··on A significant amount of programming is done by superstition
A significant amount of programming is done without articulating either the problem or the solution.
mwsherman··on The importance of user interface, illustrated by the Go flag package
The flags package is one of the few Go stdlib’s where I haven’t been won over; I felt that the cost of my understanding the abstraction was higher than just implementing the logic.

https://github.com/clipperhouse/gen/blob/master/main.go#L67

mwsherman··on A scheme to lower engineers’ wages
These are not engineers being “added”. They are being allowed to cross borders.

What it does is equalize wages. They (outside the border) make more, and we (inside the border) make less.

mwsherman··on A scheme to lower engineers’ wages
There’s no doubt the promoters of the bill are self-interested, and one should read the fine print, but: the notion of excluding workers that are on a different side of a border, in order to “protect” wages, is deeply immoral.

Do the wages of the people outside of the border count? Theirs would be going up, due to a more liberal immigration policy. The idea that such a bill would “lower engineers’ wages” seems to be selective in which engineers we are talking about.

mwsherman··on How Betty, Who Is 89, Gets Her News
Seems that Readability/Pocket/Instapaper go a long way toward that goal. Safari has a reader view too.

So perhaps it’s simply a browser or setting to always use the Reader view.

And, fwiw, Medium is doing essentially this. Perhaps ads, which hurt usability in obvious and non-obvious ways, are enough of a detriment that they will trend down. Or, perhaps web ads will be more like TV commercials, where nobody mistakes them for the main content.

mwsherman··on Fear of Failure and Lack of Speed in a Large Corporation
Another way to think of this: the likelihood of “new Y” being invented and succeeding is small, in any circumstance. The circumstances that lead to Y’s existence are very contingent. The default is for Y to fail, or not to exist in the first place.

The established company became so because they made X succeed, which itself was unlikely. So what are the chances that the same organization would also make Y succeed? Multiply the probabilities.

So yes, it’s politics and incentives, but we must also beg the question a bit further upstream.

It’s why I marvel at questions like, why didn’t Microsoft invent the iPhone. The better question is, why would it?

Tim Lee lays it out nicely here: http://www.forbes.com/sites/timothylee/2012/05/27/two-views-...

mwsherman··on The joyless world of data-driven startups
“Epistemology”, he whispered.

Using data is good, but “based on” offers a lot of wiggle room. A 10% increase in CTR is nothing to sneeze at, but it does not answer the question of the best use of your engineers’ (or designers’, or marketers’) time. Should they have instead been working on the thing that has a 50% chance of a 20% improvement? How do we account for all the data we didn’t bring to bear?

The data is small, the interpretation is big.

There is also the problem that philosophers call “regress”, which is that every rational decision has to trace back to a premise that one assumes in. Should we be in the business we are in, compared to all the other potential uses of talent? We can’t know that empirically, at root.

mwsherman··on As Big as Net Neutrality? FCC Votes to Kill State Imposed Internet Monopolies
This may be helpful, but seems to miss the point. Municipalities are the ones who created the monopolies.

The wired networks were conceived as utilities during their initial buildout, thus the monopoly. It takes years of lobbying and dealmaking for a new ISP to enter a municipality (see Google). That's unrelated to the state, in most cases.

mwsherman··on FCC Passes Strict Net Neutrality Regulations on 3-2 Vote
This argument is the essence of ad hominem. First, that disagreement is unreasonable. Second, that those who disagree have ulterior motives.
mwsherman··on Machine intelligence, part 1
“40 years ago we had Pong”

That’s very slow progress. We still have games on 2D, pixel-based screens. Better ones, but same qualitative idea; same order of magnitude in human experience. (VR is has not proved mainstream, yet.)

I say this because real-world, qualitative change is really slow. Despite all the activity in dating or real estate sites, most marriages and houses look like they did 40 years ago.

I say this to point out the difference between technical progress and real-world change. It takes a lot of the former – many orders of magnitude – to move the latter a few % points.

mwsherman··on Email Encryption Software Relies on One Guy, Who Is Going Broke
The question is the wrong one. The better question is: why does so much software choose to depend on an underfunded library?
mwsherman··on The legacy ad industry is wrong about ad blockers
I’ve never seen that, but of course possible.
mwsherman··on Syntax across languages
Fun read. A nit, if the author is reading:

C# allows Unicode letters as identifiers, so the variable regexp should reflect that. This might be handy: http://www.regular-expressions.info/unicode.html#category

Also, would love to see Go in there. Similar identifier regex.

mwsherman··on Never trust a corporation to do a library’s job
s/corporations/organizations
← PreviousPage 4 of 14Next →