ageitgey

There's a note on the README about false positives from virus scanners on Windows, so maybe that is all it is. But I can't speak to it beyond that. I've only used the Mac OS version.

(2022) - I don't think there is anything new here since this was announced two years ago.

FWIW, OpenFPGA on the Analogue Pocket works pretty well. Many of the most popular MiSTer cores were ported over and there are some nice desktop tools to make it easy to configure [1].

[1] https://github.com/neil-morrison44/pocket-sync

Yeah, I'm not sure what the joke is but I don't get it. It's just a really long video of 'training' a cat to play piano that is very obviously fake/cgi.

Average home electricity prices are 0.29/kwh in London. The street lamp charging costs 0.24/kwh, so actually slightly cheaper for most people than charging at home.

By comparison, a Telsa Supercharger will run about 0.65/kwh here (but is obviously faster) and other brands I've seen as high as 0.85/kwh for rapid charging.

However if you agreed to an electricity plan where you pay floating hourly rates and you can charge at home with the right equipment, you could in theory pay less by only charging at low rate times overnight.

But given the the limited amount of driving I do (basically just weekend trips), the street charging costs are usually so low as to be a rounding error so it is not a big deal either way.

In London, many neighborhoods have added charging sockets to the existing streetlight posts. I park on the street and can charge overnight easily and cheaply just by parking on the road and plugging in.

I'm not saying this will work perfectly in every location or for all cities. But it is relatively easy in many areas to add charging ports with existing infrastructure.

For the lazy who just want to know if this makes Python faster yet, this is foundational work to enable later improvements:

> The initial benchmarks show something of a 2-9% performance improvement.

> I think that whilst the first version of this JIT isn’t going to seriously dent any benchmarks (yet), it opens the door to some huge optimizations and not just ones that benefit the toy benchmark programs in the standard benchmark suite.

It's not that you would replace a normal solar battery system with your car. It's that a car you already have would be able to automatically help with power balancing in emergencies or automatically reduce your power bill when they are plugged in anyway without you doing anything.

https://octopusev.com/ev-hub/what-is-v2g

For example, if there's an excess in wind energy at a given moment, the system can automatically funnel that energy into cars attached to the grid instead of turning off turbines, saving everyone money.

Whenever there is a new technology, people get caught up on how the new thing doesn't work exactly like the old thing and get hung up on small differences. With online shopping, it was "it will never work if I can't try on clothes". With iPhones, it was "it will never work without a keyboard." But in the end, the new thing does new things which creates a new market.

Electric cars have all sorts of awesome features that ICE cars don't have. You don't have the same quick refueling but you have lots of other advantages. They can act as a giant battery for your house. You can keep them charged at home with no effort and never go to a gas station again. If there's a gas shortage, you can still power them easily without waiting in lines. If you live in a 220v country, you can literally just plug them into a wall anywhere if you run out of power or want to charge overnight. Most electric cars let you turn them on and heat them up remotely with no need to leave the engine running and with no fear of generating fumes in an enclosed space.

While the article is about companies who post job options that aren't actively being hired, there's another worse issue going on with actually 100% fake job listings.

Right now, there is a big issue with fraudsters pretending to be real companies and posting fake job listings in their names with fake contact info. Some job boards make this way easier than it should be. The "employee" responds and even goes through "interviews" with a fraudster, only to be asked for banking information or ask to pay fees or something like that. Then the fraudster disappears, leaving the "employee" thinking that the real company cheated them. The real company wasn't involved at all.

Be careful out there if you are looking for a job. This is a very common scam right now. Verify contact information with the actual website of the company, not from a job posting.

I'm not making any suggestions about your case. I'm just curious based on your experience, would it be plausible for someone in England to just sign up with a GP in Wales if they wanted to if the treatment offered was different than England (or the reverse)?

I'm not suggesting that anyone commits fraud or suggesting that this was a practical solution in your case. I'm just curious if that is illegal or how that is enforced.

All I could find was this: https://www.england.nhs.uk/ourwork/part-rel/x-border-health/...

The Eurostar is great and I've taken it quite a few times. But realistically, it is way too expensive right now to seriously compete with flying for most people. It is typically 2x-3x as expensive as flying from Heathrow/Gatwick to CDG, for example. That really adds up if you are traveling with multiple people. We really need more services and more competition.

> This seems to be in the same spirit as audiophiles claiming they can hear the difference between various speaker cables, or the "hints of dark chocolate" in wine tasting.

I can see why it would seem like that if you aren't seeing it, but it's not the case. The differences in color banding are pretty big if you are on a screen where you can see the background shading clearly.

The brightness of your monitor and the relative brightness of your room will matter a lot. In a bright room, you might not be able to see the subtle banding in the background of the images. But if you are looking at a bright monitor in a dark room, the difference is very obvious.

Your viewing environment will matter a lot. In a dark room with a bright monitor, the banding in the background of the example images is pretty bad (if you are looking for it). But if you have a laptop in a bright sunny room in front of a window causing back lighting, you probably won't be able to see it.

> The major advantage of AA/CP is they are simple screen projection protocols capable of being supported for decades.

There's a whole weird gray market for Chinese-manufactured "computers on a stick" that reverse-engineered the CarPlay protocol and use it to display a custom computer/Android interface in your car. You can do things like stream Netflix, use Chrome, install any Android app, play games, etc.

I'm not recommending it as it seems both sketchy and unsafe, but it's a funny niche that exists. For some reason, they are usually marketed on AliExpress as a "Car AI Box" but of course have nothing to do with AI.

Here's a random example video of what they look like: https://www.youtube.com/watch?v=LrFHjrYnoDE

Once you've owned a car with a decent HUD, it really does feel insane to go back to a car where you have to look away from the road for something as basic as checking your speed. It's a great feature.

The funny part is that while VW's bad software has been the main negative review of their otherwise great EVs, the general perception online is "Carplay/Android Auto works great, so who cares?"

I own a VW EV and use Android Auto with it. It's great. I would not have bought the car if it didn't support that. The real miracle is how well Android Auto/Car Play integrate with the car itself - speakers, microphones, steering wheel controls, heads-up display for directions - it all feels like a native experience.

> it's been incorrectly marketed as autopilot instead of a "SomeCoolName Cruise", consumers get the impression that it's superior to other self driving features

I don't think that's quite a fair view of history.

BlueCruise and SuperCruise are both limited systems that only promise hands-free driving on pre-approved sections of highway. That's all they are marketing and all they are aiming to do right now. It's not anything like full self-driving in city streets and not being sold that way.

Tesla's Autopilot has very directly promised full self-driving capability in city streets for years but failed to deliver on it. That's completely different than what BlueCruise and SuperCruise are promising.

For example, in 2019, Elon Musk promised that you'd be able to "for sure" rent out your current 2019 Tesla in the year 2020 as a fully autonomous robotaxi:

“From our standpoint, if you fast forward a year, maybe a year and three months, but next year for sure, we’ll have over a million robotaxis on the road,” Musk said. “The fleet wakes up with an over the air update; that’s all it takes.” [1]

In hindsight, he was wrong and didn't deliver that. You can decide if the Tesla is just over-ambitious on timelines or if their plan can't/won't ever work. But you can't blame this on the name of the product. Tesla has been following a very different self-driving product strategy than Ford and GM (and several other automakers) and so far that plan hasn't paid off.

[1] https://techcrunch.com/2019/04/22/tesla-plans-to-launch-a-ro...

N64 carts only go up to 64 megabytes, not gigabytes.

Not to pick on the original post, but one of my all-time biggest data pet peeves is map diagrams that aren't normalized by population/some kind of denominator. They all end up as slightly janky population maps with weird anomalies that only make sense if you know the true denominator anyway.

To your point, Kew Gardens and Richmond Park both show up with blank hexagons on the map. But Kew Gardens doesn't allow bike traffic while Richmond Park is full of bikes.

Side note: As someone who lives by Richmond Park and visits often, I wouldn't be surprised if there are more bike vs. deer accidents than bike vs. car accidents.

There's some videos up:

https://youtu.be/yY4w-bIH3X4

You have three options:

1. Legal

2. Financial

3. Personal Relationships

Unless you are also a founder with a significant stake, you probably only have option #3.

First, you have to know the legal ownership structure of the company. If the founder owns majority control of the company, there's not a lot you can do on the legal side.

Second, you need to know where the money is coming from. If it's all from investors, you might have a little leverage but not much. If you all go to the investors and say you all want to kick out the founder, lots of bad things could happen. The investor could get scared and pull out (especially if they mainly have a relationship with the founder). The investor might side with you but still not have a way to kick the founder out other than threatening not to invest additional money. Maybe the investor decides the drama isn't worth it and walks away completely. It just depends on the the situation and ownership agreements.

Third, you can just go talk to the founder. This is probably your best bet (like others have said). Maybe you can come up with a way to convince the founder that they did a great job of getting you to where you are, but their skills are better spent starting another new company while the team carries on with this phase of growth and development. Tell the founder that they are really good at zero to one, but you need to carry on from here more systematically to be successful.

But no matter what seems fair or who did how much work, you probably can't just "kick out" the founder. You'd do best to figure out how to make a deal where everyone hopefully comes out reasonably happy and the founder can save face.

If that doesn't work, all you could really do is quit en masse and go start something else without the founder.

Definitely oldies now! Their hits are two decades ago. If you were born in the 80s, Linkin Park is as old now as the Beatles were then.

If you haven't played with recent video tools, you might think this is a "someday in the future concern." But the current off-the-shelf tools for transforming one actor into another are mind-blowing: https://www.youtube.com/watch?v=eIJXOU83fqE&t=0s

And that's just what's available off-the-shelf, not what specialized studios are doing with huge budgets. The Marvel movies are already close to being "animated movies in a live-action style". Even basic helicopter shots flying over real-life locations like the coast of California are totally CGI because it's easier to animate than get a permit. Actors entire bodies will be replaced on a whim because they didn't like the costume choice. Putting a new face on someone is not that big of a deal.

I'm a CTO of a Series A start-up with open dev roles posted. Right now, my inbox is completely overwhelmed with emails from outsourcing companies offering to do the work. So based on that data point, I'm guessing there is a lot of competition for general outsourced work right now.

Part of this might because a lot of new start-ups got into this space during 2020-2022 acting as a light middleman between overseas devs and US-based companies while not providing much besides facilitating payment. So you have all the traditional outsourcing companies fighting it out with all the newer companies in an economy where spending is way down overall.

All that being said, it seems like great experienced devs in Eastern Europe are still getting snapped up quickly.

In England, the procedure costs for the entire healthcare system fit in one Excel sheet: https://www.england.nhs.uk/costing-in-the-nhs/national-cost-...

He was heavily involved in the design and development of Deepsea Challenger (along with Ron Allum, the lead engineer).

There is a lot more backstory here.

When James Cameron designed Deepsea Challenger with Ron Allum (his sub designed to go to the deepest known point on earth) he was competing against another team. For structural integrity under repeated compression/decompression, he used steel for the hull of his vessel. The other team was using a composite hull like OceanGate was using. James Cameron raised strong engineering concerns against using a composite hull and told the competing team that they would die if they used that technology due the the risk of the hull shattering under repeated compression/decompression cycles. In the end, Cameron won the challenge by diving to the lowest known point on earth and the other sub was never used. He went much, much deeper in his sub than the Titantic site (which he has also explored many times).

So when OceanGate lost their vehicle, it was basically (in Cameron's view) confirming what Cameron had said all along. It would be like if you had been warning for years that building houses out of straw was a fire risk and then someone's straw theater burned down.

James Cameron's point was that there has been over 70 years of engineering work done on deep sea subs to make them safe and that the OceanGate team was ignoring that knowledge and taking unnecessary risks.

Would be really cool in Spanish!

Here's my take on launching a business like this. Feel free to ignore it if you disagree because I could very well be wrong:

These kinds of tools are potentially very useful and the underlying tech is pretty neat, but they are marketed the wrong way. Normal users want a problem solved, not a tool. And there's several tools like this now. This one may be way better than the others, but a consumer will just see them all as equal competitors.

You could sell the same system into a professional market that has to manage a lot of documents for a lot more money. Go talk to someone who is overwhelmed by documents (lawyers, law enforcement, auto insurance, construction, whatever industry you think has money and an acute problem) and figure out what solving this problem actually looks like for them. Then sell that solution. It could be as simple as re-packaging the existing solution but explaining how it solves a specific problem for a specific type of user.

Presumably because OpenAI charges per-request fees. These services aren't training an in-house model, but are leveraging an LLM via an API and they have to pay those fees.

(I'm not saying this to imply that building a service like this is simple just because an API is used - there's still a lot of plumbing work and designing a way to manage document embeddings, etc).