Hacker Newsnew | past | comments | ask | show | jobs | submit | jwommack's commentslogin

The M chips make for real bad developer experience when trying to support interoperability.

Lots of docker images and other virtualized software just can’t be emulated like it used to on them. We’re how many years in on this issue?

When most of your customers won’t be on a high priced Mac that’s a real issue and it forces businesses off the platform.


I can run x86_64 docker images just fine, fine enough to reproduce errors I see in prod at least.

I'm using Colima along with Rosetta for Linux.


There’s a reason they stopped selling Photoshop in non-subscription when the pricing difference was something like this.

$699 vs $49/mo. Now maybe $20/mo since they no longer force bundling and have lower cost feature restricted versions.

The math was easy at the time because you’d break even in 2 years with the old price. Now it’d be like 4 but plenty of users were on CS2 for 6 or so.

Since 12x is your 1 year break even rate it’s not too unreasonable. Especially if you’re not upgrading it ever year.


Windows updates on older Eee PCs got/gets brutal a few years in. Mostly due to them needing to take up more and more hard drive space.

I remember there was a significant security update a few years back that would just constantly retry on Eee PCs because it needed like 20Gb of storage it would basically never get. Made it essentially unusable.


They pretty much already had the "Python 3" issue with PHP 6, the unicode PHP


Yeah, but it never came out. A bullet dodged.


They've been doing that since 5.2


The primary beneficiaries of this are basically Netflix and Google, no?

My understanding is most of these policies have always been targeted at the large streaming services that are putting the most burden on the networks and who often already have special interconnect agreements.

The claims that these policies would be aimed at small businesses seem largely to have not been realistic.

The whole situation with lying about what you’re ad spend is getting you seems like a much bigger small internet business issue.


> My understanding is most of these policies have always been targeted at the large streaming services that are putting the most burden on the networks and who often already have special interconnect agreements.

"large streaming services that are putting the most burden on the networks"?

That is completely non-sensical. It is not streaming services that are creating a burden, it is the ISP's customers. The ISP's customers are asking for the bits. They are using their Internet service / connection to get what they want (video). Is HN 'causing' the traffic to flow over my ISP's pipes when I reply to comments, or is it me (the ISP customer) when I click on "reply"?

The streaming services are not dumping bits on to the network like a chemical plant polluting a river. The streaming services are sending bits that the ISP's customers requested.

If the ISPs can't deliver the traffic that their customers want then they need to architect their network to handle it.

ISPs are selling access to the Internet, streaming service are on the Internet, and so ISPs need provide it or stop advertising that they're providing Internet access.

The streaming services are paying for their Internet connection, and ISP's customers are paying for their Internet connection, and it is the job of the ISP to connect the two.


But we (the ISPs) made assumptions about aggregate consumer bandwidth requirements back in the early 2000s and now they're cutting into shareholder value! Would someone please, PLEASE, think of the shareholders?!


There's simply no way for any ISP to allow all customers to use their full bandwidth, that just doesn't work, both technically and with the prices consumers pay. When building (and replacing parts to add capacity) the network ISPs look at current usage, average usage of each customer, account for future growth, etc, results in some cost that ends up as part of what customers has to pay. If traffic increases much more than expected then the upgrade/investment made that was expected to last X years would last years less, so a new upgrade is done which results in increased price for the end customer.

If the new investment was paid in part by Netflix and others like them, as they drive most of the increase in bandwidth, then the cost would be pushed to Netflix customers through Netflix.

Either way, end consumers pay what is needed to upgrade the network, otherwise the network is congested or the ISP goes out of business.

I would argue that ISPs should not have these massive profits, if they do then they are overcharging their customers. And the fix for that is to have actual competition, consumers must have multiple choices, which for the most part the US does not have in the ISP market as you have no sharing of last mile access and block municipal infrastructure at many levels.


> There's simply no way for any ISP to allow all customers to use their full bandwidth, that just doesn't work, both technically and with the prices consumers pay.

I know how over-subscription works.

But tough shit. As an ISP customer I'm paying for $x/month for x bps. I've cut a cheque now provide me with what you advertised.

> If the new investment was paid in part by Netflix and others like them, as they drive most of the increase in bandwidth, then the cost would be pushed to Netflix customers through Netflix.

The streaming services are paying for their Internet connection, and ISP's customers are paying for their Internet connection, and it is the job of the ISP to connect the two.


Do you think its feasible to build a network that is not oversubscribed? Do you think the current cost of Internet subscriptions could pay for a network that is not oversubscribed?

Netflix and others for sure do not pay a lot for their Internet connection, they do not really have one in this context. Most of the traffic comes from their CDN nodes, these are place directly inside ISPs own network (a server with multiple 10Gbit connections), they do not pay the ISPs a lot for this as its also a large benefit for the ISP since otherwise this traffic would come over peering instead and would have to travel further as CDN nodes can be geographically distributed.

The TCO of the ISPs network investments and operation is paid by their end customers, the current low cost is only possible if their network is oversubscribed.

All these things are simply solved by having a competitive market, countries that have that pay much less in Internet subscription costs. Competition is enabled by municipal networks (the ones here do not offer their own Internet service, they put down fiber when other stuff needs to be put in the ground, then let ISPs connect to their network), or by forcing ISPs to share last mile connections to other ISPs.


If the ISP has over-subscribed their network that is not the ISP customer's problem, and that is not the problem of the (e.g.) streaming service.

The ISP either has to reduce their service speed or charge more.


>ISPs are selling access to the Internet, streaming service are on the Internet, and so ISPs need provide it or stop advertising that they're providing Internet access.

How long until Comcast advertises its proprietary Comcast Network?


> How long until Comcast advertises its proprietary Comcast Network?

AOL / CompuServe here we come (again). Walled gardens for everyone.


If ISPs promised not to try to get into the business of competing with those service providers this would be a stronger argument.

When this arose, it was partly a lever to make their own offerings more attractive.

Turns out they mostly sucked at those offerings and have shut down or spun them off so it's less urgent right now, but don't assume it won't happen again!


What's the phrase? You'll own nothing and be happy.


Things of possible note:

* This is a survey published in 2018 with (survey) data gathered in two months of 2016

* mobility data comes from a different source for each country and they vary wildly in the period covered. Notable Italy and US data is post 2008 and the others pre 2008. Looks like there’s as much as a 12 year difference in child earning data across countries.

* sample characteristics (the income question mentioned in other posts, not clear these are used for quintiles):

* US 20k, 20-40k, 40-70k, 70k+

* UK £1.5k/mo, £1.5-2.5k/mo, £2.5-3k/mo, £3k+/mo

* FR (UK ranges but €)

* IT <€(?)1.5k/mo, 1.5-2.45k/mo, 2.45-3.35k/mo, 3.35k+/mo

* SE (SEK) <33k/mo, 33-42k/mo, 42-58k/mo, 58k+/mo

* population samples per bracket differ most between US and EU, EU comparisons are pretty similar. Nearly 2x income bracket 1 pop and sample % in EU(.27-.33) vs US (.16/.18)

* From what I’ve found 2004 (UK data) £ conversion rate to USD was ~1.9. (2012)€ was ~1.34. USD and Italy are from 2011-2012, a 2012 $ is 0.82¢ in 2004 $s. So 70k=57.6k and £3000/mo=04$68.4k=12$83.1k, €3000/mo=12$46.8k

* age dist differs non-UK EU has fewer under 30.(~.2 vs .27)

So the lowest UK rate is ~2x the lowest US rate, the US rate is a smaller portion of the pop/sample, and the US upper rate is nearly 1.5-2x of the sample as the Non-US (.39 vs .17-26). The EU top rate is ~0.56 the US top rate when converted to $ using 2012 numbers.

There’s also another important dynamic here. 2011-2012 was the European debt crisis. Italy, in particular, saw a big unemployment spike over the course of the selected years.

[the study](https://pubs.aeaweb.org/doi/pdfplus/10.1257/aer.20162015)


The bottom end of that range is $125k/yr. That’s within the average of a senior software engineer. It’s not a big leap or, really, hard to get in that context.

The article specifically sites this as for senior software roles.


How come websites like payscale, glassdoor, etc say 125k is what Google pays in Bay Area?


The Stanford study here cites an 11% reduction from ~8% to ~7%.

This is purely for surgical masks, it found no impact from cloth.

By far most mandate compliant masks even in high mask populations are cloth. There might be a conflation going on here.


Consider applying for YC's Summer 2026 batch! Applications are open till May 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: