shimon

This actually seems like a structural weakness in cloud, where so much of revenue comes from the relatively few customers who are very deep in their usage. On the other hand, you can usually rely on close relationships with those customers to identify their significant needs. However, it might be harden to implement them without the bench of deeply knowledgeable engineers.

I studied CS and now work in software systems for biomedical research. It's difficult to overstate how different the fields are, so I don't entirely agree with this statement. But I do agree there are going to be lots and lots of huge discoveries in biology in the 21st century.

The main difference is that CS attempts to generate and study complex systems built from well-understood components, whereas biology attempts to understand and manipulate systems that evolved naturally over eons.

Imagine dropping a fully functional internet-connected Google Home Hub into 1960-era humanity and asking them to figure out how it works so they modify it to sound like Walter Cronkite. There are thousands of problems on this order of complexity in biology. It's wild.

If you could isolate everyone in the whole world for 3 weeks, you would cut down on a lot of infectious disease as most individuals would beat their infection in that time and little transmission would occur.

However, you still wouldn't fully eliminate infections because some people would not beat their infection in that time. They might sustain an infection for longer because of their own health, e.g. they could have an immune disorder or be on immunosuppresant meds.

Of course you can't actually do that because (1) the whole world's human population is incapable of coordinated action on this scale, (2) there would be lots of very valid exceptions -- e.g. are you going to split up a family with 10 kids or kick very ill people out of hospitals?, and (3) lots of people would try to cheat and many of them would succeed.

Basically we should be aiming for maximal vaccination, good treatments, constant monitoring, and good non-pharmaceutical interventions (e.g. ventilation, masks in crowded places). Lockdown-type measures should be used when there is the possibility that critical healthcare infrastructure could fail under excessive load, i.e. things like childbirth and broken legs could become dramatically more deadly.

Presumably association including temporal association around events that are described by others. If I text <shadow friend> frequently around the time I post about frisbee games and bluegrass concerts, odds are they are engaged in those types of activities with me.

The point is the tactic of giving the first number, though successful in ending many unappealing recruitment attempts, also has the unintended side effect of making it impossible that you'll ever get an offer SIGNIFICANTLY ABOVE your target range. Because comp boosts in the 50-100% range are possible, and you actually want to see those, you should let the recruiting org offer the first number.

These typically reference model codes that are often proprietary. You may still be able to get them from a library but if your question is along the lines of "how many receptacles can I put on one circuit" you need to look in something like the International Residential Code[1], not in your city/state statutes.

[1] https://www.iccsafe.org/products-and-services/i-codes/2018-i...

Yes, the modern solution to this consists of a pressure-reducing valve (refills when pressure in the boiler system drops) and a backflow preventer (for the safety issue).

This idea that DIY wiring can invalidate your insurance policy is widely repeated, but I don't think it's true.

Homeowner's insurance routinely covers preventable disasters. Did you leave a candle on or a heater plugged in and start a fire? Did you not cut down a tree that was looking unhealthy and it fell on your home? These are all situations where your insurance would pay a claim. The cases where they would not pay a claim are more along the lines of "deliberate arson".

You may well need to certify that work is permitted in order to get an insurance policy.

Relevant reddit thread: https://www.reddit.com/r/HomeImprovement/comments/7vwe23/doe...

This is true for most organizations. Note though that the practice the author describes at Amazon is that everyone gets the document to read at the beginning of the meeting. The structure of the meeting means that (1) someone is forced to produce the narrative ahead of time and (2) everyone has to read it alone together.

That's an odd setup, but it's probably what makes this system work -- an organizational admission that this meeting setup is actually better than walking people through slides. One advantage the author doesn't mention is that this "alone together" mode gives each participant more time to think before having to fit their thoughts into a group discussion, which likely promotes divergent thinking and gives people room to ask important, difficult questions.

I signed up a few months ago and had this type of validation as well. I didn't get a postcard, presumably they hadn't started that practice yet.

This is a great perspective but it's important to remember your sample bias -- you're only making judgments about the offender you've identified. Or more snarkily, you've only caught the people dumber than you.

When you're talking about basic attacks to steal some money, this is a substantial but not world-changing effect. On the other hand, a government or cartel willing to invest significant resources in cyber offense can really move the needle. What we've seen Russia, North Korea, Israel, and others achieve is probably only a fraction of what they've actually achieved.

Not OP, but there are some common problems, assuming they were a non-founder CTO:

- Instead of leading an eng team, maybe CTO meant "our entire tech team".

- Maybe the founders sucked, or had personal conflicts.

- Maybe the business was broken in ways a technologist can't fix.

- Maybe the business was viable but this individual did not enjoy being a part of it. Could be anything from "there are too many salespeople yelling into phones all around me" to "the technical problems were not the sort I am interested in / good at solving".

I strongly agree that capital flight is a huge and under-recognized driver of high asset values in the US (particularly equities and real estate).

But how do you even begin to estimate the rate of this effect, or how much further it might go, or what asset classes it might affect? For example: this was likely a major force in the surge in Bitcoin value almost a year ago, but that asset has since collapsed; will a similar effect take place in other assets?

I assume that OP meant saying no to taking on more expenses / ongoing obligations than you need, not funding reserves.

The best examples are probably in diet, e.g. Weight Watchers, and Jenny Craig. There's a huge industry around drug & alcohol rehab, though it's not easy to see which parts could support startup-type scale. A bunch of these needs are served in the healthcare industry, often with much room for improvement. Healthcare is a totally massive industry.

It's not an easy place to build a business, but there are lots of resources spent on these problems now. They affect families, businesses, and individuals. Figuring out how to line up these things and make a viable business is complex, but some people will figure it out.

What's your rationale for this? Beyond interest rates, which are clearly the result of more aggressive, far larger-scale easing by central banks than we've ever seen before.

This is a great description of what Waymo is doing, and why Honda/GM etc. don't want to become a commodity provider (thus letting Waymo capture most of the value).

From Waymo's perspective, that's exactly the scenario they want. Focus all your resources on developing a tech advantage that others can't match, and have others in the ecosystem serve the lower-margin stuff.

If this sounds familiar, recall Google Search and Web Content.

You have too short a time horizon on your "nobody".

This approach would seem to negate the whole "convenience and simplicity" motivation behind using Amazon Prime.

I think you're underselling this. You're right that malice is relatively rare. Ambiguity, however, is the norm.

For example, say you have a quick discussion and decide you will do task X and your colleague will do related task Y. You nod and head off to work. For most people, this feels like a commitment but if they haven't explicitly promised a deadline then there's no assumption of when it'll be done. So it might actually never be done, and whatever commitment has been made is super weak.

But then: you write an email "Thanks for the chat. I'll do X this week and you'll follow up with Y in time to ship by Thursday the 11th". By getting explicit about the timing you're really asking them, in a specific but not confrontational way, to seriously consider the timing portion of the implied commitment. Maybe they're super swamped with something else and assumed you'd be OK waiting? Maybe the work is more complicated than you realized and will take longer?

For most people, in most work environments, these kinds of ambiguous commitments are more common than clear, thoughtful ones. But anyone in an interaction can help steer it toward clearer commitments, and this is frequently very valuable. It's valuable to you because you know what you can really expect from others (or if they let you down, you know the difference between them failing to magically read your mind vs. do what they promised). It's valuable to others (as long as you're not a jerk about it) because it'll help clarify misunderstandings before they become a bigger problem.

You're not wrong, but you do realize that the outcome of a midterm election won't affect the president's authority to sign or veto bills?

The value proposition of this product is claiming to do too much. Because I don't have a pre-existing belief in the value of integrating all these functions in a single "PIM" (what is a PIM anyway? is my smartphone a PIM?) I'm stuck wondering how this might help me, and reluctant to try something that might not be valuable until it has replaced a huge set of habits/tools.

If I want to benefit from trying this PIM, where should I start? In case it helps, I'm a gmail/gcalendar/emacs org-mode user.

Just trying to put some reasonable bounds on this estimate. I assumed 12 weeks is "the holiday shopping season" where Amazon has a lot of extra workers who would benefit. And I guessed 250k of them getting $3 raises. You could also guess at the total cost in other ways, many of which would be superior.

The overall goal of these lazy estimates is to understand how significant a cost this is for a company like Amazon. Could activists have hoped for more? Are other companies, for whom the costs would be felt quite differently, likely to offer something similar?

I'm not making the argument that this is peanuts. Maybe it is, maybe it isn't. I'm just trying to get some perspective on the scale here, relative to other things vying for the time and attention of Amazon's management.

Quick back-of-the-envelope math:

Let's assume this raise will affect 250k workers this coming holiday season, bumping their pay from $12 to $15 (+$3).

Over 12 weeks at 40h/week, this represents an additional $360M in labor cost.

Using AMZN's 2017 revenue of 177.9B, this equates to 17.7 revenue-hours.

What's the wrong here?

Waiting tables is a hard job, and not everyone can do it well. It's physically and mentally taxing, and to make money at it you have to be charming which is harder to train than, say, packing boxes in a warehouse.

Good-paying waiter jobs tend to be in high-cost cities, so you'll spend more money on housing, transit, and food than a worker in a suburban warehouse. Waiters are also harder to substitute for - due to the small scale and personal nature of restaurant service, it's unlikely a restaurant could drastically improve waiter productivity without impacting the customer experience.

Also, to your point in comparing the hourly rate: warehouse and office jobs offer a different type of financial value to workers - relatively predictable schedules. If you're a student, working a few evenings a week and raking in tips is great. If you're trying to support a family, getting consistent full time hours may be more appealing than a higher hourly wage with lower/less consistent hours.

What you're describing is a workplace with a generally broken approach to meetings. The best thing you can do with standup in this environment may be to eliminate it, or minimize it as you've described. But there is often a place for a standup meeting in many work environments, and barring the other stuff I'd suggest before lunch is a good time - anchored to another interruption that would happen anyway, and provides an incentive to finish quickly without making it a hard stop.

Not that you asked, but this sounds like a totally dumb time for this meeting.

It creates a reliable interruption, such that if it happens you're likely to be distracted enough that it isn't even worth trying to get back into serious work for 45 minutes before you need to go home.

It's also late in the day, so any tasks that come up must pass a high bar to be actually remembered and resolved. A task must either be so urgent that it is worked on immediately (and then people will resent the standup for making them stay late) or so important that it can be remembered the next day. This means a whole bunch of simple-to-fix issues are probably being ignored, either by forgetting or just not being raised in the first place.

I'm not surprised that the meeting often doesn't happen. A better time for a standup is before lunch or earlier in the morning.

Lisps are great. Why do you think they haven't been at the forefront of any big trends in application development? Things like the web, mobile apps, data science ...

My guess at an argument here is that the languages popular in the 80s drove the curriculum design of most Computer Science education, and the relative absence of the Lisps (to today) makes the languages seem less approachable to practicing programmers than they really should be.

On the other hand, you can find a lot of Lisp's influence in something like Python (though obviously with many differences both superficial and deep). So in that case, why are Python IDEs so much worse than what you'd see in Lisp? (And is that even the case? Maybe there's just more Python devs and thus more IDEs and like anything, most are crap; but if there one or two great ones then does Lisp really have an advantage there?)

Maybe the reason for this has more to do with the type of development structure that helps a language get scale.

The rich powerful development environments you've described exist primarily in proprietary, integrated environments. If you want to integrate the editor, debugger, OS, and language, it helps to be able to coordinate the design of all those components.

On the other hand, languages that have gotten to huge popular scale have typically been more open in their specification and implementation process. Perhaps this is because the creators of breakthrough tools that drive language adoption (like web frameworks or data science kits) prefer these tools, or because the sorts of conditions that lead to creation of such tools are inherently marginalized ones. In other words, if you're a happy iOS developer using a lovely integrated world of xcode and swift, you're not going to spot a dramatically underserved development niche.