Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
dmazzoni
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
15 ms
·
151.
▲
by
dmazzoni
3y ago
Seven figures means a million dollars or more. A large fraction of us on Hacker News are in the California Bay Area / Silicon Valley, New York, Seattle, Chicago, Toronto, London, and other major cities where a million dollars gets you
152.
▲
by
dmazzoni
3y ago
It's Chrome OS. It runs a Linux kernel but by default it's designed to be a very simplified, locked-down OS that runs the Chrome browser and Android apps. It also has a built-in Linux development environment that you can enable, w
153.
▲
by
dmazzoni
3y ago
First of all, you're off by an order of magnitude. Second, I don't think it will be that long. There are already LLMs as good as GPT-3 running on average laptops and even phones. In the next couple of years, you'll see: - Ord
154.
▲
by
dmazzoni
3y ago
ALso, AIs will get much better at delegating. Currently ChatGPT doesn't know it's bad at math, so it can convert a story problem into an equation better than a human but then mess up the arithmetic or forget a step in the straight
155.
▲
by
dmazzoni
3y ago
This might be more convenient than braille in a few circumstances, like replacing a braille menu in a restaurant - but it definitely doesn't make braille obsolete. Also, note that many blind users access Be My Eyes using a refreshable
156.
▲
by
dmazzoni
3y ago
They're not getting rid of that! This is a complementary service, there are still plenty of scenarios where a human makes more sense.
157.
▲
by
dmazzoni
3y ago
This isn't trying to replace human volunteers, but complement them. I know blind people who are making just as many human Be My Eyes calls as they were before, but they're using Be My AI even more, for things where they wouldn
158.
▲
by
dmazzoni
3y ago
Trolling isn't a big problem, but many humans have good intentions but just aren't that helpful. Some want to be chatty or give unsolicited advice. Some just aren't very good at tech. Some just aren't very good at descri
159.
▲
by
dmazzoni
3y ago
That death is of course tragic, but it's important to put it in context. That's the only death of a self-driving car. That's the only one. There was a safety driver with their hands on the wheel, and they didn't see the
160.
▲
by
dmazzoni
3y ago
No, this is genuinely a useful technology. Blind people are getting a ton of value out of it. It's way faster than humans, you can do it quietly without bothering anyone, and honestly it's better at describing some things than 90%
161.
▲
by
dmazzoni
3y ago
And that's just one more thing that blind people don't want to have to deal with when they just want to go about their lives. Humans are slow and unreliable. AI is fast and consistent. Both are imperfect. You shouldn't rely o
162.
▲
by
dmazzoni
3y ago
Being blind isn't necessarily dangerous at all. Lots of blind people have their homes set up with everything carefully arranged and no hazards. They prepare food, clean their house, commute to school, to jobs, and to appointments via f
163.
▲
by
dmazzoni
3y ago
Also, there are many low-stakes situations where AI is actually quite a bit better than the average human: 1. Take a photo of a piece of tech: AI can not only describe what's going on but tell you exactly how to fix it (e.g. "it&#
164.
▲
by
dmazzoni
3y ago
Yeah, but the alternative right now is completely random, unverified, human volunteers with no qualifications and no training. While the vast majority have the best of intentions, some people are just...not that good at helping. I've b
165.
▲
by
dmazzoni
3y ago
Have you ever needed to do something like make an appointment but kept putting it off because you just really didn't want to talk on the phone? That can happen to anyone. Some blind people are introverts and don't want to talk to
166.
▲
by
dmazzoni
3y ago
Fun fact: Cisco is actually headquartered in San Jose, but I believe that both companies have large engineering offices in both San Francisco and San Jose. For those who don't know, they're both part of the same metro area (Bay Ar
167.
▲
by
dmazzoni
3y ago
It's a widespread issue. With DALL-E 2, yes, the vast majority of the time you do get undefined blobs full of mistakes. You can get some amazing things, but only with clever prompt engineering. This is about DALL-E 3, which is just ann
168.
▲
by
dmazzoni
3y ago
It's still a great data set because it's real-world, useful, high-quality, and an excellent size (~60k examples), and the "correct" classification isn't very subjective at all - humans can easily get 99% accuracy.
169.
▲
by
dmazzoni
3y ago
That blog page is not showing state-of-the-art results, it's just taking relatively naive SVM implementations and comparing them. The original research paper that introduced the MNIST data set was getting around 98% accuracy, and today
170.
▲
by
dmazzoni
3y ago
Employees are only human. Even smart, savvy, well-trained employees can be fooled by good social engineering every once in a while. The key to good security is layering. Attackers should need to break through multiple layers in order to get
171.
▲
by
dmazzoni
3y ago
Anything's possible, but the simplest explanation (per Occam's razor) is just that the employee was fooled. Is it plausible that if a good social engineer cold-called a bunch of employees, they'd eventually get one to reveal
172.
▲
by
dmazzoni
3y ago
If my math is correct, 15 grams would mean more than 20 normal boxes of Sudafed, if each box is 24x 30mg. I usually buy boxes of 48x 30mg, but even then I'd have to stash 10 boxes for it to be illegal.
173.
▲
by
dmazzoni
3y ago
Is the behavior that a running .js script is fully loaded into memory and the file doesn't need to exist documented, supported behavior? What if, hypothetically, the system was suspended in the middle of script execution, and the resum
174.
▲
by
dmazzoni
3y ago
Yes, which is why the radiation limits for cell phones was set to 10x lower than the lowest amount that might be theoretically even slightly harmful. Now one testing lab in France says that one particular phone is slightly (less than 10%) h
175.
▲
by
dmazzoni
3y ago
It's affecting lots of phones. "Smartphone radiation tests have so far led to 42 sales stops in the country. It is the first time Apple has been affected by such a move." From https://www.reuters.com/technolog
176.
▲
by
dmazzoni
3y ago
Yeah, but do you really think any of that applies to things like the Linux kernel? You think with thousands of talented developers they have problems as simple as duplicate code?
177.
▲
by
dmazzoni
3y ago
Just in case anyone doesn't already know, you can do that with a data uri as well: data:text/html,<h1>My%20small%20website</h1><p>Look,%20it's%20real!</p> You can use a data uri generator to
178.
▲
by
dmazzoni
3y ago
This reminds me of the Tank neural net urban legend: https://gwern.net/tank
179.
▲
by
dmazzoni
3y ago
I think Python was popular as a general-purpose language first. After all, there was a reason people put so much effort into writing Numpy in the first place. I think a lot of people were attracted to the language design, as captured in the
180.
▲
by
dmazzoni
3y ago
Years ago when I saw a paper on that topic, I tried recording my own keyboard and trained a ML model to classify keystrokes. I used a SVM, to give you an idea of how long ago this was. I got to 90% accuracy extremely quickly. The "gues
More ›