HN user

mikewave

215 karma
Posts7
Comments85
View on HN

This is a fantastic use of the D-Wave technology and really highlights the ability to use annealing for a wide range of use cases. I'm definitely going to follow this development closely!

It will be interesting to see how this works on the Advantage2, especially when it moves past the prototype stage. The greater connectivity between qubits may allow for an even better demonstration.

I've been seeing Mac folklore posted to HN recently - this is a fantastically well produced podcast that goes through folklore.org and other articles from the time period, with asides, tidbits, interviews etc. that help paint a picture of this fantastic era in computing history. Well worth a listen.

Well, if your system elastically uses GPU compute and needs to be able to spin up, run compute on a GPU, and spin down in a predictable amount of time to provide reasonable UX, launch time would definitely be a factor in terms of customer-perceived reliability.

Over the past five years, there has been undeniable hype around quantum computing—hype around approaches, timelines, applications, and more. As far back as 2017, vendors were claiming the commercialization of the technology was just a couple of years away—like the announcement of a 5,000-qubit system by 2020 (which didn’t happen).

Inaccurate. D-Wave did in fact launch a 5000+ qubit system named Advantage in 2020.

I don't mind that idea - if it turned out installing heated seats failed in say... 28% of installations, fine - charge me more for a model with functioning heated seats.

I know you just chose this as an example, but clearly it wouldn't be acceptable to anyone to have a failed installation of something like this. On one extreme, it could cause a fire; but otherwise, it's just a button that doesn't work, and damages the quality reputation of the car.

Processor binning only works because it is a black box, indistinguishable from the outside. You can purposely disable components if your yields get high enough, etc. and nobody will ever notice the difference.

For vehicles, automakers will learn from customers that this kind of nickel-and-diming is not appreciated, when people turn to other manufacturers that still "get it", e.g. Mazda.

Same here. It's barebones, but it's also cheap and plentiful on the used market, has enough functionality to be useful without being overwhelming, and fits easily in a backpack. Did some noodling around with it connected to my iPhone on a plane recently, no drivers needed if you get that Lighting-to-USB adapter, phone powered it fine, and I was able to drive Korg Gadget and make some shitty techno.

Some of us would love to work on a team like this. It would be nice to have the option. Your definition of "acceptable" might not actually result in teams that can take on the big challenges we face as a species as men who did find this kind of thing acceptable retire out of the workforce.

Correct. The availability of cheap credit + higher than normal resale values for vehicles of late has led to a situation where people are overextending themselves and buying vehicles that cost 80% or more of their yearly income in order to "keep up with the Joneses". I know some people in this situation, and their ability to afford $600/mo payments (plus insurance, plus gas, plus maintenance) is leading them to actually do less: fewer outings, fewer trips, less expenditure elsewhere, perhaps no summer camp for the kids, etc.

America badly needs a return to the much lighter S-10 and Ranger of the past, with a 4-banger and a small back seat. These behemoths with V8s are not necessary for most people, and neither is the level of temporary luxury they convey. Let's face it: after a couple years your shiny new truck is just as noteworthy as every other truck on the road, which is to say, most people don't care about it / aren't impressed by it anyway.

Yes, if you have H.Pylori it can cause all sorts of problems, acid reflux, bad breath, gastritis, ulcers, etc. I'm no doctor but I have been tested for this in the past; I think it was just a blood test, nothing invasive, and it's curable.

Better to get it sorted sooner than later. That GP visit can also be a good time to ask about all sorts of other small issues, and a blood test is a good idea at least every year or two. Ask about hormone levels, iron, vitamins, everything; for you it's just a lab visit.

I can't wait until it's possible to get these tests a few times a year and have all the values charted somewhere tied into the personal health ecosystem. Ironically, as someone interested in online privacy, I find the medical data the least concerning; I'm willing to let all of that go if it means I can get back some good answers or better healthcare. I can understand why some people aren't, of course, and they should always have the right to it; and it's good that this medical industry privacy situation is influencing other industries as well. But still! Imagine if some ML system with access to the data of a million people like me could show me exactly what I should stop eating, or exactly what supplement I need more of....

I'm no farmer, but I'm all ears! If you have dry useless soil and can't water it, does it not have some residual value as an export product?

I suppose this is assuming that there's not going to be water in the future, and that you just want to extract whatever use there is left from an area before giving it over to desertification.

Every kind of functionality you can imagine already exists in open formats with powerful programs designed to make it easy for you to compose rich documents.

Not only that, but this kind of thing has existed for ~30 years now. I remember MS Word 6.0 on Windows 3.1, 16 bits and 4MB of RAM, with large documents with significant styling, embedded Powerpoint and Excel documents, MS Draw vector art, etc.

I concur that the line-based approach most Git tooling uses is not good enough. At the very least, even for code, it should be integrated with Tree-Sitter etc. by now and be comparing changes at the AST level, not just the raw text; and when it comes to documents, we should get side-by-side diffs showing the rendered document.

But is there a market for this, besides just you and me?

There are real, usable, near-realtime Quantum Annealers, which are different from gate model quantum computers, but have proven to be much more amenable to physical implementation. These are more geared towards optimization functions, which can have practical use for logistics, physics simulations, and more.

Gate model computers interest everyone because of the potential to factor large numbers; this could potentially render a lot of existing and past cryptography open to decryption that otherwise will be locked up for aeons, especially considering the apparently decaying effect of Moore's Law of late. However, it probably takes millions or billions of gate qubits to do that, whereas thousands of annealing qubits are enough to solve some interesting problems, and to contribute to finding better solutions when combined with classical compute in a hybrid setup.

Yes, have done DEB and RPM and Docker and Makefiles with `make install` and IMO the differences for a packager are pretty minimal and not worth having a holy war over.

Anyone smart enough to manage one format can probably manage any of them.

a pretty bad way to write and manage packages

Cool, good to know this nearly 30 year old conflict is still just as active as ever. Tens of thousands of working DEB packages notwithstanding.

This is just it. I have an amazingly smart home: in every single room, I have installed a robust remote control for lighting that cannot get lost, can be replaced for $5 if it breaks, and is perfectly customizable with interchangeable face plates. The kids can even paint them!

I don't need my lights to turn on and off on their own. I don't buy that anyone does, short of perhaps someone who has a physical disability for whomst it would be too difficult. For $20 I can get a timer - perfect for the closet light we always used to leave on by accident.

I think this is basically conspicuous consumption, tasteless pseudo-luxury. I am not impressed by your IoT colorful lightbulbs; impress me with your conversation.

Worst part about discontinuing the Focus is that the Focus Electric went with it. Sure, it was a regulation-requirement car, but it drives well and does what city folks need; 100KM range is enough for most people's day. Works well as a second car.

We desperately need sub-20k electric cars and they just don't exist.

Let’s get some housekeeping out of the way at the top: DevOps isn’t a job. I see too many articles written about ‘DevOps Engineers’ and ‘DevOps Specialists’ and my understanding of the term doesn’t allow this kind of terminology. DevOps is a design goal like Agile.

Sorry, but the sheer number of people out there working in jobs with "DevOps" as their title and the thousands and thousands of positions that recruiters are trying to fill at the moment means that this idea, which you're far from the first to posit, is simply not correct anymore (if it ever was).

A DevOps is a sysadmin who can program. That's it, end of story. This puts these people in contrast to the non-programming sysadmin (e.g. perhaps of the Windows era of expecting a proper UI to configure things), and the non-administrative developer (again, also very much of the Windows era of producing enterprise software that is "thrown over the fence" to Ops).

The Unix administrators who scripted things in the past are the direct spiritual predecessors of the modern "DevOps" who simply have newer and better tools to accomplish the same goals.

This means a number of tools can be DevOps as a service.

No, that doesn't make any sense at all. Tools merely provide capabilities; these must be exercised by people who understand them in order to realize value. Even as basic things are automated away, there is still a considerable role for the people who are able to knit systems together, especially for being able to unblock pure-Dev teams (for which there is still a tremendous role), facilitate test engineering and QA needs (again, a specific skillset that shouldn't always be lumped upon developers), and so forth.

Put differently: when you want to take a shop to the level of maturity to have proper automated deployment, or you get stuck with some broken CI problem you cannot figure out, when you need to make your existing product multi-region, when your org acquires some other org and you need to bridge the products together, when you need to expand to a hybrid or multi-cloud system... all of these kinds of situations require the kinds of skillsets that are labelled as "devops" today.

Don't confuse the map for the territory.

The new Grafana alerts do absolutely nothing to help with this.

I'm at the point where I would pay 5 figures a year for something purely to do better alerting inside or alongside Grafana. Clicking alerts together is a nightmare when I have a ton of identical systems I need to configure. Same for dashboards - the limitations of the current mechanism are too severe.

I'd build my own templating mechanism for it, but I still want the alerts visible in Grafana itself. Zabbix has the power to do all this but with a UX that is not ideal....