Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×

Submission + - What are the FLOSS community's answers to Siri and AI? (upon2020.com)

jernst writes: A decade ago, we in the free and open-source community could build our own versions of pretty much any proprietary software system out there, and we did. Publishing, collaboration, commerce, you name it. Some apps were worse, some were better than closed alternatives, but much of it was clearly good enough to use every day.

But is this still true? For example, voice control is clearly going to be a primary way we interact with our gadgets in the future. Speaking to an Amazon Echo-like device while sitting on my couch makes a lot more sense than using a web browser. Will we ever be able to do that without going through somebody’s proprietary silo like Amazon’s or Apple’s? Where are the free and/or open-source versions of Siri, Alexa and so forth?

The trouble, of course, is not so much the code, but in the training. The best speech recognition code isn’t going to be competitive unless it has been trained with about as many millions of hours of example speech as the closed engines from Apple, Google and so forth have been. How can we do that?

The same problem exists with AI. There’s plenty of open-source AI code, but how good is it unless it gets training and retraining with gigantic data sets? We don’t have those in the FLOSS world, and even if we did, would we have the money to run gigantic graphics card farms 24×7? Will we ever see truly open AI that is not black-box machinery guarded closely by some overlord company, but something that “we can study how it works, change it so it does our computing as we wish” and all the other values embodied in the Free Software Definition?

Who has a plan, and where can I sign up to it?

Comment Re:Other than Brother... (Epson EcoTank) (Score 2) 387

I use a non-EcoTank Epson, but I have added on a 3rd party Continuous Ink System. It consists of four ink tanks outside of the printer (which you refill with 3rd party ink, cheap), with surgical tubing running to fake ink tanks inside the printer. The fake ink tanks have a little chip with a sneaky-beaky reset button on it, which scrambles the serial number of the ink tanks. This works so well, and is so cheap that I don't take into account ink costs anymore. Cheaper than my laser per page, but I can print out art and sell it.

Submission + - Dissecting a frame of DOOM

An anonymous reader writes: An article takes us through the process of rendering one frame of DOOM (2016). The game released earlier this year uses the Vulkan API to push graphics quality and performance at new levels.
The article shades light on rendering techniques, mega-textures, reflection computation... all the aspects of a modern game engine.
Power

Tesla Preps Bigger 100 KWh Battery For Model S and Model X (theverge.com) 113

An anonymous reader writes: Tesla will soon offer a 100 kWh battery for the Model S and Model X that will allow for increased range -- perhaps as much as 380 miles for the Model S. Currently, the 90 kWh batteries are the company's largest capacity. Kenteken.TV is reporting that the Dutch regulator that certifies Tesla's vehicles for use in the European Union, RDW, has recently published a number of new Tesla variants. RDW's public database now includes entries for a Tesla "100D" and "100X," which are titles that follow Tesla's current naming system based on battery capacity. The listing for the 100D claims the vehicle has a range of 381 miles or 613 kilometers. The motor output is reported as 90 kilowatts (121 horsepower), which is the maximum output the Tesla motors can sustain without overheating. Autoblog notes that EU range estimates tend to be more optimistic than those issued by the U.S. EPA. A more realistic range might be 310 to 320 miles.

Comment Re:A surgeon got cancer from one of his patients (Score 1) 121

I read a while back that a surgeon accidentally got cancer from one of his surgical patients:

http://www.nejm.org/doi/full/1...

So, it appears that cancer can move between hosts in a mechanical fashion.

I found a nastier one a while back: guy has tapeworm, tapeworm has cancer, tapeworm spreads its cancer throughout the guy's body as it wriggles around.
Here's the story. One of the interesting things was that the tapeworm tumors had differently-sized cells, so they were easy to differentiate from the host's cells. Now, that's not exactly cancer being transmitted, insofar as it wasn't his cells that were turning cancerous, but they were growing/multiplying and helped cause his death. It's like being infected with some other animal's cancer.

Comment Re:Can you *know* something you don't even believe (Score 2) 385

Among epistemologists the near-consensus is that belief is one of the necessary ingredients of knowledge.

Cite? I know lots of things I don't believe in. For example, I have quite a lot of knowledge about how magic works in various fictional systems. I find it much more likely that you're mischaracterizing the belief/knowledge of epistemologists than that they're really that stupid.

I feel the same way you do, but:
http://plato.stanford.edu/entr...
"There are three components to the traditional (“tripartite”) analysis of knowledge. According to this analysis, justified, true belief is necessary and sufficient for knowledge.

The Tripartite Analysis of Knowledge:
S knows that p iff

p is true;
S believes that p;
S is justified in believing that p.
The tripartite analysis of knowledge is often abbreviated as the “JTB” analysis, for “justified true belief”."

certainly reinforces at least a classical view that epidemiology claims belief is necessary for knowledge (with the proviso that there are modern theories of knowledge that disagree.)
The kicker seems to be in the use of the word 'justified', which I think I'd characterize as a weasel word on Wikipedia.

Comment Re:Averages are misleading (Score 2) 127

For example, the average person has approximately 1 testicle.

Whenever people trot out the tired statistic that 70% (or whatever) of drivers believe they're better than the average driver, to mock the idea of how well people rate their own abilities, I trot out the statistic that 99.9% of humans have more than the average number of eyes. We assume gassian distributions with no idea if that's actually representative of the results.

With that said, at least the air force example is a case of misunderstanding of how averages concatenate, which is a slightly more complicated poor understanding of statistics. (It doesn't take long to calculate 0.3^10 and realize that it's an extremely small number, but that's not intuitive if you haven't taken statistics classes, or, in my case, lots of chemistry classes: when you have a complex synthesis that has seven steps, each with 80% yield, you realize you're going to need kilograms of starting material to even measure your final product.)

Comment Re:CNC Mill (Score 1) 165

I'm going to argue with the OP. Most retrofit cnc kits won't give you 0.002" repeatability because the lead screws are too sloppy. LEGO bricks are built to 0.0002". A Tormach 440 cnc, cost $5K, with ballscrews rather than leadscrews, still quotes positional accuracy at 0.0013". Prices rise quickly from there. The Haas minimill, at $34K, appears to claim positional accuracy of 0.0005", although it's not obvious that its repeatability is that high.

Comment Re:None. (Score 1) 165

Your best bet would most likely be a small SLA printer. Like the a Formlabs or a Autodesk Amber.

We tried printing LEGO form-factor bricks with a (now older) Stratasys mid-range SLA. The results were marginal for plates, ie topside-only, and flat-out unacceptable for the bottoms of blocks. Blocks that were correctly dimensioned to snap onto a lower block, and thin enough to snap in beside another block, broke during use. We had to make fat blocks -- increase wall thickness -- that couldn't nest in beside other blocks to get ones durable enough to survive a dozen attach cycles.

Comment Re:They can probably sell 300 (Score 1) 276

If you could reduce that 2700 pounds down to 2200 pounds, with 200 horsepower, that could be both fast _and_ nice-handling.

They could also slap 300 horsepower in there without any trouble, there are a number of V6s that would do the job in that small space.

Based on my experience, of having put a lightweight-for-a-V8 buick 215 in a little british car, and then replaced that with a 1.3 liter suzuki that weighs less than 100kg, I vastly prefer every bit of the handling characteristics of the smaller, lower-power engine. The v8 was much better at drag racing, but the i4 in the same car with the same tires completely obliterates the v8 in initial off-the-light acceleration, turning, braking, and generally everything that car does when it's not on a dragstrip. The DeLorean weighs 70% more than my car, so it won't be as marked a change, but I've become a big advocate for removing weight rather than adding power.

Comment Re:They can probably sell 300 (Score 1) 276

They're talking about putting in an engine w/ 300-400 hp, which doesn't sound like a huge number, but the original car weighed 2700 lb. If they're close to that weight with a decent transmission it will be damn fast.

I can see a few problems though. The original frame/engine mount only had to deal w/ 130 hp, so probably some modifications needed there. The weight distribution was 35/65 front/rear. With a presumably heavier engine/transmission and anything else needed back there that ratio could get problematic.

A lot of people seem comfortable driving fairly high-performance older cars with serious rear end weight imbalance, like corvairs, which had almost the same weight distribution. The PRV V6 wasn't very lightweight. If they were willing to continue with Colin Chapman's original design philosophy and sacrifice some horsepower for a lot of weight reduction, there are 125 horsepower engines that weigh under 100kg out there these days. If you could reduce that 2700 pounds down to 2200 pounds, with 200 horsepower, that could be both fast _and_ nice-handling.

Comment Re:This could rock (Score 2) 276

By the way, the fine article refers to this change in regulations which allows low-volume manufacturers to produce classics without things like airbags or OBDII. That's interesting because a ton of cars vanished from the landscape not from lack of consumer demand but because the design was unable to be economically and aesthetically updated to meet new regulations. The Jaguar XKE is a perfect example. Again, cars using modern drivetrains, fitted in classic bodyshells, would be a neat mix of beauty and durability. As anyone who drives an old car will tell you, driving with a full toolbox in the back, out of necessity, can be a headache.

Comment This could rock (Score 2) 276

The engine and transmission were underwhelming in the original, and modern engines with significantly higher power, and lower weight, are ubiquitous. If it had decent engineering -- as in, if they are using a modern differential and wheel bearings -- the stainless steel body could make for a car that lasts for decades. However, man, the dash design was seriously dated and would need a ton of updating to look even modern, much less futuristic.
In other words, I'm dubious about the commercial viability of that car's design with only a new engine. A new engine, drivetrain, interior, and electronics, would make an interesting car... but that's not quite the same as "only the engine [will be changed]".

Comment Re:Don't buy new. (Score 2) 111

If I buy a car from the dealership for $50,000 and then try to sell it 10 minutes later it's now worth $25,000 at most.

If I buy a used vehicle from someone for $15,000 and try to sell it 10 minutes later it's still worth about $15,000.

Why take the hit?

Also a car that is worn out still has a pretty poor resale value even if it is just a year old.

Maybe it's that I'm coming at this from a different usage pattern than yours, but new cars make a lot of sense to me. I bought a new Subaru in 2001. I put 230,000 miles on it, and when I got rid of it, the head gaskets were leaking, two CV joints were failing, the center differential limited slip clutch pack was no longer working, the shocks were shot, and it was way overdue for new timing belts, the cumulative cost of which far exceeded the car's value, so it went off to a junkyard. The previous Subaru I had was totaled. Before that I had a Nissan, that was totaled. I've never sold a car to anything but a junkyard for scrap value, so while the depreciation on a new car is high, it's also meaningless to me.
In contrast, when the last Soob went off to the junkyard, I bought a used Subaru, and over the last three years I've had to replace the transmission, two CV joints, the shocks, and am about to have to replace the u-joint/cv joint (I haven't taken it out yet so I don't know which) that goes to the rear wheels, none of which I had to do on the new car, because it was new. While it was busy depreciating, I was busy driving it rather than fixing it.
For my use case, new cars are significantly cheaper, have higher availability, and I don't end up spending weekends in the snow replacing CV joints.

Slashdot Top Deals

There's no sense in being precise when you don't even know what you're talking about. -- John von Neumann

Working...