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

Hyster (not Hister as far as I can tell) is a registered trademark in the US. Regardless:

"The HYSTER trademark is filed in the category of Education and Entertainment Services" [0], so you can safely ignore any demands to rename, as it doesn't conflict. Embarrassing for them that their lawyers don't understand even the basics of trademark law.

Edit to clarify as some folks here are as confused as sHyster's lawyers: Trademarks are not exclusive, they're restricted to a category or categories. You can be Apple in the category of computers, but not in music if there's already an Apple in that category (unless you have $500M to spare [2]).

  [0] https://www.trademarkia.com/hyster-77843354
  [1] https://tmsearch.uspto.gov/ for a more authoritative source than [0], but it doesn't allow deep linking
  [2] https://en.wikipedia.org/wiki/Apple_Corps_v_Apple_Computer

> Embarrassing for them that their lawyers don't understand even the basics of trademark law

Or more likely do, but hope that you don't...


I find it harsh but fair. The second paragraph reads:

"Firms like Jeppesen have offerings in this space, but Scikit-decide, together with a narrow- and wide-body fuel consumption model built by a professor at the Delft University of Technology and wind data from NOAA, offer an open source solution."

So this is being positioned as an open source alternative to commercial offerings, which clearly it is not. If instead he made the limitations clear from the outset, that this is just a demonstration of what he/the tools are capable of and that this is not in any way intended for real-world use, it would put the rest of the article into better context.

I also don't see how Scikit-decide being developed by Airbus gives OP any domain authority. I use open source tools from all sorts of commercial ventures, and know nothing about how they themselves use them, or what for.


> you have to use the Latch app to open your apartment door

So what happens if you arrive home with a dead phone?

And possibly worse, they have a log of when you come and go?


You have a unique code, don't they already have a log of your coming and goings?

But I would absolutely not install an app for that.


They do. But the apartment complex gets a nice little kickback, especially since an app for unlocking your front door is one that has at least a plausible argument for having your location, which means a whole bunch of location data that can be mined/sold/both.


Usually smart lock does not have network connection(due to using battery).

But the app on your phone does not have that limitation.


I am a student and where I live most student housing has an app to open the door instead of a normal key. My current place used Bluetooth, the previous one was even more inconvenient because it used NFC. It sucks, and if you accidentally forget your phone at home you're fucked, but the landlords do it because it lets them block old tenants from entering without having to switch the locks.


> And possibly worse, they have a log of when you come and go?

Most apartment buildings in the US already use face-tracking cameras to get this log


> like you said, its on the verge of ripping itself apart

They didn't say that - they said "the (exclusive speciality alloy) parts are literally ripping themselves apart". That sounds like a seriously flawed design to me


It's just hyperbole.

When parts spin that fast centrifugal force becomes a serious problem. Everything is a spring, nothing is actually 100% solid. Anything you make that spins that fast will at minimum grow larger in the direction of centrifugal force and shrink in perpendicular axis. If you don't account for that stretching and shrinking things like turbine blades will end up scraping the walls of the enclosure at speed.

Even minor variations in weight get multiplied and create exponentially higher stresses on the part. If you don't have the right alloy with the right hardening/annealing parts will rip themselves apart from attempting to support their own weight against the centrifugal force.

There's a reason modern high-bypass turbofan jet engines use monocrystaline blades because growing a single crystal of such size is the opposite of easy or inexpensive. Plus in the core you are dealing with so much heat no alloy exists that can withstand it for long so you have to machine cooling channels inside the blades (which have complex curves) to keep them from getting soft and flying apart.


Yup, you nailed it. This is exactly what I meant when I said “tearing themselves apart”


So not literally ripping themselves apart then


I'm not aware of an aftermarket head unit that can't connect to CAN bus with the appropriate adapter (which are widely available).


Got into a weird situation recently where my cheap android head unit was displaying Car Play from my phone, but (Google Maps) audio from my partner's.

Directions weren't coming from my phone's speaker or the car, but testing audio in the Maps app did (from the car).

Still not sure what combination of connections it had managed to get itself into!


> There are some countries where the truth is not a defense against libel

Germany, for example. Utterly bizarre and baffling that a democracy protects its politicians this way. /s


It's important for people to be able to critize elected officials.


> they don't have "memory", they have no capacity to learn outside of the training stage

The underlying LLMs don't, but the agent frameworks around them do.


The underlying LLMs do, but we choose not to use the capability because it's expensive and doesn't quite work as well as we'd like it to, or quite in the way that we'd like it to.

We are perfectly capable of running LLMs in a way that does a backward pass to update some or all of its weights after every user message. But, naively implemented, you only get partial, fragmentary absorption of the info in those messages, it costs three times as much compute, and you lose out on the ability to implement a ton of optimizations that making modern LLM serving economical.

If you want to do it, though, ask your friendly neighborhood robot to get it working with a tiny model (whose full precision weights fit several-times-over on your machine's resources).


While I agree it’s theoretically possible, do you believe this capability exists in the LLMs we use today?


Depends on whether you define "the LLMs we use" as the collection of weights or if your definition contains the software stack that runs it

Technologically the LLMs we use today don't implement this behavior, but you could take the weights of Sol and add a couple (very large) patches to vllm (or whatever OpenAI has today) and have a version of Sol that does have "memory"


"Dave constructs a homemade megaphone using only some string, a squirrel, and a megaphone"


There's a ton of experimentation on it, the field is called continual learning. It's not something you need to believe in like Jesus, you can just go read about the current state of things.


And they invariably should not. Or, should, but, should leave it out until it sucks far less.

That's been my experience, anyway. Memories from 100 prompts ago tainting what I'm trying to do right now.


I've recently done the same for a Bluetooth LED matrix display, then took it a step further assembling my own hardware (ESP32 controller and RGBIC matrix) to have control over the firmware for features that weren't possible with the off-the-shelf device.

None of this would have been possible (or at least practical from a time-value perspective) for me to do without Claude driving the whole exercise.


Using a sound card (read the comments)



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

Search: