Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
wongarsu
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
30 ms
·
751.
▲
by
wongarsu
8mo ago
Not in terms of people printing lego bricks. But at least as an adult, designing things in Fusion and printing them scratches a similar itch as building lego. And 3d printing is now pretty accessible to the 14+ age group. I doubt this will
752.
▲
by
wongarsu
8mo ago
> I heard your same rant in the 1980s The two options would be that either the perception is unsubstantiated but persists, or there has been a continuous decline for the last 40 years. I'm strongly leaning towards the latter. I also
753.
▲
by
wongarsu
8mo ago
On the one hand, not publishing any new models for an architecture in almost a year seems like forever given how things are moving right now. On the other hand I don't think that's very conclusive on whether they've given up
754.
▲
by
wongarsu
8mo ago
The results would probably be underwhelming. The bitnet paper doesn't give great baselines to compare to, but in their tests a 2B network trained for 1.58bits using their architecture was better than Llama 3 8B quantized to 1.58bits. T
755.
▲
by
wongarsu
8mo ago
Approximately as challenging as training a regular 100B model from scratch. Maybe a bit more challenging because there's less experience with it The key insight of the BitNet paper was that using their custom BitLinear layer instead of
756.
▲
by
wongarsu
8mo ago
I've also always though that it's an interesting opportunity for custom hardware. Two bit addition is incredibly cheap in hardware, especially compared to anything involving floating point. You could make huge vector instructions
757.
▲
by
wongarsu
8mo ago
I take "every character" to mean "anything that was represented in a reasonably common pre-unicode code page or character encoding, as well as anything that might come up in OCR output of text documents". Emojis obviousl
758.
▲
by
wongarsu
8mo ago
Because the scraper is either impatient, careless or indifferent; and if they scrape for training data they don't plan to come back. If they don't plan to come back they don't care if you tighten up crawling protections after
759.
▲
by
wongarsu
8mo ago
As long as the inverter feeds at most 800W. That's about 4A, on a circuit designed for 16A. You need a new meter, but the old analog meters can run backwards and you can continue using it until you power company replaces it. You do ha
760.
▲
by
wongarsu
8mo ago
Solar on the balcony railing is very popular right now, mostly amongst apartment renters
761.
▲
by
wongarsu
8mo ago
Signed into law in 2002, with the last reactor going offline in 2023. Depending on how you count we got rid of it a quarter-century ago. Not the best decision, and a major reason why Germany uses so much coal and gas today. But outside some
762.
▲
by
wongarsu
8mo ago
GPT 4 was rumoured/leaked to be 1.8T. Claude 3.5 Sonnet was supposedly 175B, so around 0.5T-1T seems reasonable for Opus 3.5. Maybe a step up to 1-3T for Opus 4.0 Since then inference pricing for new models has come down a lot, despite
763.
▲
by
wongarsu
8mo ago
> The problem is that once you start letting an AI "expert" dictate tone and logic, you effectively lobotomize the writer’s original intent Isn't that what grammarly has always been, since long before the invention of the
764.
▲
by
wongarsu
8mo ago
sound great
765.
▲
by
wongarsu
8mo ago
You will produce output that emulates the patters of Shakespeare's works, but you won't arrive at them by the same process Shakespeare did. You are subject to similar limitations as the llm in this case, just to a lesser degree (y
766.
▲
by
wongarsu
8mo ago
For certain types of "human in the loop". If it can't write working code without a human in the loop then it's not AGI. But a human-level coder also has lots of humans in the loop: a more senior developer doing code revi
767.
▲
by
wongarsu
8mo ago
For a single container the difference depends on your exact setup. I typically run docker in the same WSL instance as my other wsl stuff, so starting a single docker container adds 0 additional VMs, while starting the container as a wsl ins
768.
▲
by
wongarsu
8mo ago
So the fire department routes the call to a nurse who decides how urgent the issue is and dispatches ambulances according to that. And that nurse is employed by the same company who operates the ambulances. I can't imagine a clearer co
769.
▲
by
wongarsu
8mo ago
running docker on windows runs docker in wsl (which is a HyperV guest). The project offers running docker images as wsl instance (which is a HyperV guest), no docker involved
770.
▲
by
wongarsu
8mo ago
At least the DOM APIs are ostensibly designed to work in multiple languages, and are used by XML parsers in many languages. Some of the newer Web APIs would be difficult to port. But the majority of APIs have quite straight forward equivale
771.
▲
by
wongarsu
8mo ago
Even if we completely ignore the artistic and technical merit and complexity (which I would argue mostly moved to post-processing images) or the ability to catch or create the moments you want to capture, simply the fact that someone has to
772.
▲
by
wongarsu
8mo ago
The ability to run docker containers as wsl instances looks nifty. A bit more overhead since they now run as a VM instead of a container, but you get the ability to start a shell with Windows terminal or access files with Explorer (and thus
773.
▲
by
wongarsu
8mo ago
Because ASI is 'superior to the best human' while AGI is 'matches or surpasses humans'. It's hard to agree on what 'matches humans' actually means. Would matching humans in coding involve the level of the
774.
▲
by
wongarsu
8mo ago
AGI is so nebulous we will never be able to tell if we hit it. We have hit human-level abilities in some narrow tasks, and are still leagues away in others. And humans have so vastly different skill-levels that we can't even agree what
775.
▲
by
wongarsu
8mo ago
Employees are the ones with the real power to make this hurt. The customers switching over are easily offset by the DoD contract. But losing talent over this, and having a harder time to attract future talent? That could hurt them Sam proba
776.
▲
by
wongarsu
8mo ago
The adversary aren't the model vendors, the adversary are the students. The students will modify the prompt, ask models to rewrite text in an atypical style, or use specialized services that attempt to hide the typical AI patterns. And
777.
▲
by
wongarsu
8mo ago
To add to this: another sign of posting on online boards is starting your comments with "I agree" because otherwise the other person might default to assuming you are disagreeing (as is the norm for replies), leading to a comment
778.
▲
by
wongarsu
8mo ago
You can detect if texts from a year ago used AI based on statistical patterns. Nobody is taking issue with that. But once you tell people "we will run these tests to detect if your future submissions are using AI" you create an ad
779.
▲
by
wongarsu
8mo ago
How exactly do you envision this going wrong? The rack itself won't hold electrostatic charge, and if the devices themselves want to be grounded they can be grounded through their power supply And wood isn't that easy to get to
780.
▲
by
wongarsu
8mo ago
Depending on the author 17th century English can also be very close to modern English. A couple phrases will be off and the spelling is different, but most of the difficulty is more the author using constructions that have fallen out of use
More ›