Jack000

16 ms·

121.
by Jack000 5y ago
There are other ways of doing this even without germline editing - you could for example train a linear transformer model that does DNA phenotype prediction, then just do regular IVF and choose the embryo with lowest predicted risk of (canc
122.
by Jack000 5y ago
Trying to fix aging for existing humans is like trying to patch a binary that's already running, it would be way easier to build new humans with better source code imo.
123.
by Jack000 5y ago
welcome to the bottom of the page! You're in good company though, we have tea and snacks. Unfortunately HN users are not somehow special and immune to flame bait. This type of content would dominate the front page if it weren't ma
124.
by Jack000 5y ago
I'd agree if there's actual science being suppressed, but it seems to me there is none, just a lot of speculation and finger pointing without evidence on all sides. Despite the lack of solid evidence more than a quarter of America
125.
by Jack000 5y ago
This article is kind of self defeating - based on the article the scientists expressed uncertainty over the origins of covid, but the title is "Scientists believed Covid leaked from Wuhan lab" which seems to prove their point exac
126.
by Jack000 5y ago
Most of these are just misapplications of tech by other people, you don't really need to participate if you don't want to. I'd hesitate to even call this stuff "tech" - search engines, streaming video, online retail
127.
by Jack000 5y ago
Life is long for most, but statistically there's about a 5% chance you'll die before 50, it's better to hedge your bets imo. Also, wind surfing in Ibiza doesn't sound that fun at 60. I worked as a software dev for a larg
128.
by Jack000 5y ago
There's a simple way to ensure that you're making something that people want - just solve an existing problem instead of finding a new one. Search for your problem on google. If the top results are blogspam with ads for monetizati
129.
by Jack000 5y ago
AGI is a weird term because it seems to mean different things to different people. Generally it seems to mean "parity with human cognition", but cognition is not just one thing and some ML models already outperform humans in narro
130.
by Jack000 5y ago
IQ isn't a linear scale, so 120 IQ isn't actually all that high (90th percentile vs 50) You can get half way there with proper diet, sleep and a few years of education. That said, I think it would be a terrible experience to have
131.
by Jack000 5y ago
only hire people for easily commoditized tasks (work that's done frequently and where requirements are similar from project to project) In any other case the friction of finding the person who can do a good job is not worth it.
132.
by Jack000 5y ago
I don't have any special insight into the problem, but I'd say whatever form real AGI takes it won't be a language model. Even without AGI these models are massively useful though - a version of GPT-3 that incorporates a know
133.
by Jack000 5y ago
I think it's better to think of GPT-3 not as a model but a dataset that you can interact with. Just to give an example - recently I needed to get static word embeddings for related keywords. If you use glove or fasttext, the closest wo
134.
by Jack000 5y ago
GPT3 is a huge language model, no more and no less. If you expect it to be AGI you're going to be dissapointed. I find some of these negative comments to be overly hyperbolic though. It clearly works and is not some kind of scam..
135.
by Jack000 5y ago
One thing I don't see anyone mention is that the M1Max is the probably the cheapest GPU memory you can buy. The only other way to get >=64Gb GPU memory is with the A100 (?) which is like 20k by itself. So this would be great specifi
136.
by Jack000 5y ago
So just read the paper and it's pretty much as described in the article. In the model random events happen to individuals, a doubling of capital for a good event and a halving for a bad event. The probability of a good or bad event is
137.
by Jack000 5y ago
Unless there's more to the model than described in the article, it doesn't make any sense. It sounds like none of the individuals in this model have any agency, and just have random events happen to them? If you make RNG the bigge
138.
by Jack000 5y ago
Still way too expensive for individuals and startups imo. These A10 cards are just slightly slower than RTX3090, which makes it an easy comparison. $1/hour for the cheapest option means renting for 62 days is equivalent to buying outri
139.
by Jack000 5y ago
hi guys, just a project I made for color generation using ML. Here's a writeup on the model and data: https://huemint.com/about/ a short tutorial on how to use it: - click (generate) on the top right - click any c
140.
Show HN: Machine learning for color palette generation(huemint.com)
2 points by Jack000 5y ago|1 comments
141.
by Jack000 5y ago
The problem with "free will" is that there's no physically plausible system that can satisfy the definition. If you could re-arrange the laws of physics in any way you wanted in order to give humans free will, how would you d
142.
by Jack000 5y ago
High cognitive ability is like having a big engine in a car, doesn't do you any good if there's no traction. Then there are people who love nothing more than doing donuts in a parking lot.
143.
by Jack000 5y ago
there is a form of art called "found object art", which is pretty much exactly what it sounds like. Once upon a time, I asked my art instructor what the difference is between this art and a random object on my desk, couldn't
144.
by Jack000 5y ago
hm.. have you seen DALL-E/CLIP? It's literally a language model applied to images and text. It's also the dumbest possible way to generate images (generating tokens one at a time, down and to the right) yet it works.
145.
by Jack000 5y ago
I think it's a mistake to judge the performance of GPT-3 like this because it's not comparing apples to apples. What GPT-3 is doing is equivalent to me giving you a book in Chinese (without any further context, translation or grou
146.
by Jack000 5y ago
hey, lamp and jquery still works. I've also built stuff on actionscript 3/flash right at the tail end of the flash era, it sucks to have the floor fall from under you like that but what can you do. who knows, maybe in 10 years the
147.
Show HN: Huemint – Machine learning for graphic design color generation(huemint.com)
1 points by Jack000 5y ago|0 comments
148.
by Jack000 5y ago
I started out designing everything from scratch, but realized it's 10 times faster to just use a commercial vue/html theme. You don't need custom design, especially if your product is not design-related. It's absolutel
149.
by Jack000 5y ago
> For the businesses paying FAANG salaries, its not true, or they would be paying lower salaries and accepting not getting quite at the top of the human talent pool. Maybe, but the FAANG interview process is pretty noisy, also not everyo
150.
by Jack000 5y ago
I think Codex/Copilot is the beginning of the end for software development as we know it. In 2025-2030 the majority of "programming" will consist of feeding business requirements to AI models in the form of natural language a

More ›