urthor

15 ms·

121.
by urthor 4y ago
My favorite trick is simple. I don't "fully" automate. I partially automate, then build my workflows from these composite parts. Composition. For a start, I file all my bash scripts/notes/snippets in Vimwiki. Zet
122.
by urthor 4y ago
That's true, I'm not giving any credit at all to the fact the article isn't written by the paper's authors, and is at best a summary.
123.
by urthor 4y ago
Says a lot doesn't it. Marie Jackson makes it into the citations twelve times. Otherwise not mentioned by name. Academia: You thought capitalists were greedy for mere money? Come meet academics.
124.
by urthor 4y ago
That's arguably an even greater point than anyone realises. All these geniuses... people are giving them jobs!
125.
by urthor 4y ago
Alan Kay's advocacy for funding science is a truly eminent life's work.
126.
by urthor 4y ago
Remind me, in the history of the physical sciences, how many "groundbreaking breakthroughs," came from academics teaching undergraduates? My mind is genuinely struggling... Certainly not Curie, Einstein, Fleming worked in a hospit
127.
by urthor 4y ago
Certainly it's one of the largest issues in society today. We simply don't have much of an idea how the human hormone system works, or how it reacts to all the chemicals present in the "Standard American Diet."
128.
by urthor 4y ago
The real problem is that we spend hundreds of billions advertising food. If we spent 10% of what we spent on food advertising on large, well funded cohort studies, we'd be far better off as a species. Humanity's greatest crime is
129.
by urthor 4y ago
Dentists work 4 days a week. They have a quiet chuckle.
130.
by urthor 4y ago
There's a noticeable difference in my abilities at the start and end of the day. At the start I have more processing power and short term memory grunt. At the end I have far more knowledge loaded into cache. The first part of the ole
131.
by urthor 4y ago
Save humans time with a workflow. Save humans time. Save humans time.
132.
by urthor 4y ago
UML is hard. Any abstraction can evolve into wasted time not making the thing.
133.
by urthor 4y ago
Bottom Line Up Front and "Anything worth reading is 10% of the first draft," are the two hardest procedural skills in writing. It's just so much experience to avoid those mistakes.
134.
by urthor 4y ago
I disagree, but only slightly. The advice about rotating shoes to prevent overuse injuries, paying attention to stress bottlenecks is general. Bodies differ in certain ways, but I've heard the same story as the shoes from many intensiv
135.
by urthor 4y ago
Well, you gotta make sure only 10% of users purchase. Not 50%.
136.
by urthor 4y ago
Strongly disagree with the linked article. Individual customers, and corporate customers, are wildly different. The "buy it for life" business model works very well if you're trying to launch a mass market product to individu
137.
by urthor 4y ago
This is one of the joys of command line tools. They almost never stop working on you, and are far easier to maintain cross platform. Using the same thing on Mac at work and Linux at home is a joy. Increasingly, I actively try to use the com
138.
by urthor 4y ago
As a user, software companies are usually not producing good enough software to justify paying for the price of incremental updates ;). In many ways, I feel much software needs to transition to "be maintained," very quickly. So mu
139.
by urthor 4y ago
The entire point of "everything being a subscription" is a marketing ploy. (Also revenue. "I forgot to cancel the subscription" pays the bills like no other, see Comcast's revenue.") Subscriptions have to be r
140.
by urthor 4y ago
The issue is if they did this, what would happen is: 1) An enormous pile of cash would accumulate on Google's balance sheet. 2) The spinoffs would be subject to normal commercial rules about risk, rate of return and such. 3) The
141.
by urthor 4y ago
Decision making credibility is simply not anchored in "being logically correct." It's anchored in the ability to consider the needs of the tribe appropriately. Leadership credibility in human society is mostly anchored via
142.
by urthor 4y ago
A very good point. People trivialize the progress in this progression, because it's normalized. And yet, the incredible progress in the last 5-10 years is absolutely stunning. Compared to the other "recognized professions," t
143.
by urthor 4y ago
> The next step is to make the grid flexible. But... we can do exactly that??? Model retraining is very transparent and a very straightforward way to redirect it. The core problem is that the analogies just don't hold up. Artificial
144.
by urthor 4y ago
With those big companies, commonly there simply isn't a productive way to walk the line. Saving face and preserving "decision making credibility" means that taking advice is simply not on the cards for a large portion of the
145.
by urthor 4y ago
Sound like the solution is incredibly straightforward. Spin "the other bets" out as a separate company. Capitalize it separately. Literally just don't have it as part of the parent company. The only advantage of having the co
146.
by urthor 4y ago
A very very conservative prediction given the exact same thing happened to IBM, Sun, Burroughs, Yahoo to an extent, and infinitely many others.
147.
by urthor 4y ago
That's an unfair comparisom to gym braggarts I'd say. Unlike with working out, build in public provides a valuable source of engagement and feedback. Those are real, concrete reasons, to build in public.
148.
by urthor 4y ago
The system before peer review wasn't particularly good either. Peer review is worst system possible, except for all the others that have been tried. I'm aware that many subfields have good solutions. But, from the outside as a non
149.
by urthor 4y ago
Oh good, that makes perfect sense. I've just found the idea of "1 year of experience repeated 10 times" at a dissonance with an industry very much oriented around staffing individual projects. I agree, as long as you're
150.
by urthor 4y ago
Says a lot about the software industry as a whole.

More ›