itissid

18 ms·

211.
by itissid 2y ago
Dunno about that but There is though a laser that shoots certain parasites... On farmed salmon's bodies.
212.
by itissid 2y ago
This like protomaps? https://protomaps.com/
213.
by itissid 2y ago
Awesome!. I like the focus on a specific part instead of the whole file. This is built to address a very clear set of human problem: Working memory and distraction due to clutter. In general, using a machine to work with the human to solve
214.
by itissid 2y ago
Yeah, though I do wonder for a big model like 405B if the original training recipe, really matters for where models are heading, practically speaking which is smaller and more specific? I imagine its main use would be to train other models
215.
by itissid 2y ago
How are smaller models distilled from large models, I know of LoRA, quantization like technique; but does distilling also mean generating new datasets for conversing with smaller models entirely from the big models for many simpler tasks?
216.
by itissid 2y ago
One common theme I see in the paper(e.g. in protein folding) is: "Identify what properties are important (geometry, algebra, topo) and which one is an useful prior and then "use" the guide to select an initial struct. This is
217.
Making leetcode editorials a bit lesswrong(leetcode.com)
2 points by itissid 2y ago|1 comments
218.
by itissid 2y ago
... and possibly a bit more enjoyable to read
219.
by itissid 2y ago
Barret mentioned that: If one bribes the president in appointment of an official - like an embassador - since the appointment of the embassador is an official act, under this ruling, one cannot bring this as evidence to the jury in a crim
220.
by itissid 2y ago
Could you summarize a few nuggets of wisdom about why you like wikis? Specifically, what particular ways of use are the most effective within a company?
221.
by itissid 2y ago
One issue that traditional AI Agent reasoning built on Theory of Mind(ToM) (ToM: I know what you are thinking, so I can predict what you will do next.) was hard to get right was because the domain model and interface was quite tough to desi
222.
by itissid 2y ago
Given its 4$ per month[1] and they are a non-profit with just 55M$ revenue. How is GPU cost and hence their with OpenAI for this going to work if it becomes really-good and 100's of millions of kids start using it? Renting a 24GB VRAM
223.
by itissid 2y ago
YOLO is amazing. When I first looked at the loss function in YOLO v1, I was stunned: You can do regression loss(Mean square error) instead of classification loss(categorical cross entropy) on class prediction probabilities, and it can still
224.
by itissid 2y ago
I think these companies are trying to solve a different problem: They are just inventing a way to clean streets using near autonomy. I am trying to figure out how to solve a pain-point which can work along with current or these technologies
225.
by itissid 2y ago
Thanks. I hadn't heard of them. But it would be really cool to connect with them and learn about some crossover learnings I am sure I will encounter down the line.
226.
by itissid 2y ago
Thanks for the advice! I am trying to augment street cleaning and change the alternate side parking to make it better. It happens throughout the big cities in the dense urban parts like NYC, Jersey City, SF, LA. Mine is a small city in huds
227.
by itissid 2y ago
Thanks a ton. That is great advice for strategy!
228.
Ask HN: Robotics projects in consort with the City and build advice
2 points by itissid 2y ago|9 comments
229.
by itissid 2y ago
This. Medallion systems in NYC were gamed by a guy who let people literally bet on its as if it were an asset. The prices went to a million per until the bubble burst. True story
230.
by itissid 2y ago
The problem is the transitive closure of chess move is a chess move. The transitive closure of human knowledge and theories to do X is new theories never seen before and no Value function can do that, unless you are also implying theorem pr
231.
by itissid 2y ago
MITs MO is to try to create a breakthrough tech and rapidly spin off startups from the tech. But I think the main barrier is that to operationalize hardware requires massive capital investment upfront for testing and then regulatory approva
232.
by itissid 2y ago
https://archive.ph/vElvl
233.
CoComelon Captures Our Children's Attention(newyorker.com)
1 points by itissid 2y ago|1 comments
234.
by itissid 2y ago
It's not just mapping events in space and time, it's also bringing in appropriate context and expectation of future (goals, intentions) into the present, other people's mental models into our prediction. I am not sure how abs
235.
by itissid 2y ago
Interesting. It seems most of these task target a very specific part of the brain that recognizes visual patterns. But that alone is cannot possibly be the only definition of intelligence. What about Theory of Mind which talks about the pro
236.
by itissid 2y ago
I don't think they sell it like Google or Samsung. For example Apple does not have a location intelligence team dedicated to driving revenue for store brands or targeting users that go there using precise geo location data. Google and
237.
by itissid 2y ago
Does this mean that it could be a good practice to release the auto encoder that was trained on a neural network to explain its outputs? Like all open models in hugging face could have this as a useful accompaniment?
238.
by itissid 2y ago
Look at JEPA and Modulo-LLM. Also AGI is a poor term to use because we as humans have no notion of what general intelligence is, does GI have morals and ethics, does it make decisions like we do based on executive functioning or does it wor
239.
by itissid 2y ago
That two letter tool at meta for profile access?
240.
by itissid 2y ago
Lowest common denominator to solve all these problems is either 1. Invent super cheap and scalable energy solution. Like Fusion. We could pretty much do whatever climate solution we want. 2. MMT to achieve (1) and a combination of one of: a

More ›