thesimp

In NL I can buy a base Macbook Air with 16G memory and 512G SSD voor 1199,- inc tax.

I just looked up my M1 receipt: in 2020 I bought a Macbook Air M1 with 16G memory and 512G SSD for 1399,- inc tax.

I did not expect the price for a base machine to go down in 2026.

I'm going to go the other way: if 7 days is not enough then you are not thinking hard enough about how you want to solve the problem that you have. Because if you have to do so much, type so much code, organize so many things, just to get your problem resolved, and it cannot be split up or delegated, then you are doing it wrong.

I do not know how it is for other people but I cannot be in a creative and problem-solving mood 8 hours a day. Heck: I have the best days were I actually type very little but still get to _finally_ understand and solve whatever I'm looking into.

Working efficiently is underrated.

Please the small but significant typo: it is B instead of M for the numbers.

I'm not a professional analyst but the numbers look weird to me. Prediction was to sell 507,000 in Q4, they delivered 497,570, but only made 459,445. That means that they had almost 35.000 cars "laying around" from Q3 that they delivered in Q4. That seems a bit weird to me. But maybe someone that monitors the big 3 automakers in the USA can comment on how many cars waiting in the pipeline is a normal amount.

Also the total 2024 production shows something very interesting: Tesla is a one-trick pony. They sold 1,77 million cars, of which 1,70 million were model 3/Y. So the Model S, X, and cybertruck only make up 85,000 cars. That is not healthy for the long term. Because the Model 3 and Y compete in the middle class price wise, so nobody buys expensive Tesla's. And nobody can buy a cheap Tesla because there is no 25k Toyota Corolla equivalent.

It will be interesting to see in the next years if Tesla can make a dent in other price segments.

As someone who bought Microsoft Flight simulator via Steam I concur. I swear that Flight Simulator updates use at least 3 layers of redirection between starting via Steam, getting updates for the main program, and then getting updates for the scenery files,and then getting another set of updates while already running the main Flight Sim program. Absolutely bizar.

I started with a Psion 3 and later a Psion 5MX and now 30 years later I still remember how good the user experience was. The agenda, de contact database, the way you could make your own database with data that you wanted to lookup while on the move, it all just worked. At some point in time I even had a mapping tool, I think the precursor to what is now TomTom.

Personally I think that today an only iphone/macbook with just the default software (mail/contacts/calendar/etc) comes somewhat close to the same user experience. A Windows machine, while having a million different programs available at your fingertips, does not feel as a cohesive user experience.

Dyson spheres are a nice way to look at what humans are capable of in the future. That is why they probably capture the imagination.

I actually do not see them as much more than beefed up solar panels. 50 Years ago a solar panel was something that most people saw on a novelty pocket calculator. 30 years ago people started thinking about solar panels to actually power a house off-grid. 20 years ago people started things about solar farms that could give useful power to the grid. And now solar panels are ubiquitous and just about anyone can put them on a roof of a house and the world makes them in such large numbers that it is hard to imagine the number of panels made per second.

Maybe in 20 years time we will find out that a solar panel in space is much better because the sunlight does not have to travel through the atmosphere. So why not put large panels in space. Then in 40 years time we understand that we need to position them closely around the sun to get the maximum power output. Of-course the power cord is a bit long to earth so you then also put your factories on those panels. Which is no problem because launching large reusable rockets in space is a problem that is being solved right now. And breaking factories down in container sized chunks to transport them was already invented 60 years ago with containers and container ships.

I'm actually shocked that a company that has spent 25 years on finetuning search results for any random question people ask in the searchbox does not have a good, clean, dataset to train an LLM on.

Maybe this is the time to get out the old Encyclopedia Britannica CD and use that for training input.

Interesting that they move to touchscreens. With all the Hacker News stuff that comes by every day I can easily say that 99.9% is not what I do in daily life. But building control rooms for large industrial installations is. And from my perspective we see very little requests for touchscreens. They have a place, for example on a machine in an environment were a keyboard would be in the way. But only if the operator would not be wearing gloves. Very specific applications.

For large modern control rooms users will have either a keyboard and/or a custom button panel next to the mouse/trackball. With the mouse/trackball being the primary input device to click on objects on the screen. In many cases they don't even type in values but click on buttons on the screen to ramp up or ramp down a process value. Then they don't need a keyboard.

The article suggests that ChatGPT slows down because it is winter and in the winter people (and data on which ChatGPT trained) sometimes slow down: the dark/gloomy/depressing winter feeling. To which I made the statement that ChatGPT does not know that it is winter so the hypothesis in the linked article does not hold up.

Several people have commented that ChatGPT does know the current time and date, so maybe there is indeed some truth in the linked article. But it does show how inconsistent ChatGPT can be: in any interaction I had with it I could not get CHatGPT to admit that it knew the current time, date, or season.

In order to have the "winter blues" you must know that it is winter. And in all my questions to ChatGPT I never had positive feedback that it knows what time it is, what date it is, and what the current season is.

My theory is that someone left a DEBUG=1 flag somewhere in the code and that the debug.log is filling up to 4 GB. I'm only joking a bit, I've been bitten enough times by these type of issues to know that they must happen all over the place.

The problem is not keep the battery warm while parked, it's keeping the battery warm while driving. Most EVs can "preheat" themselves before you start in the morning, the smart one's might even learn the patterns and be ready when you get in at 07:30 every day.

The problem is that driving EVs in the cold costs a lot more energy. I've got a Ford Mustang Mach E for about 2.5 years now. In the winter the range that Ford claims drops by about 30 to 35%. That is a lot of range that goes missing just because the temp drops below 5 degrees Celsius.

Luckily I'm the perfect EV candidate: my daily commute is less than 50% of the total range so I can drive 2 days to the office if needed. And I can charge both at home and at the office.

The main problem that I see is that people cannot charge at home. If you are dependent on fast chargers by the side of the road you are going to have a hard time. The downtime for fast-chargers is enormous: my personal guess would be that they do not reach the 90% uptime. Which is bizar problem to have because a fast-charger and remote monitoring of the charger condition should be a solved problem by now.