itissid

24 ms·

331.
by itissid 3y ago
Also remember bad research can come out of good universities. Remember the gzip compressor beats BERT paper that showed gzip beat bert at many KNN based tasks? Or just Google for Wansink Cornell. So best to treat every paper like a i.i.d sa
332.
by itissid 3y ago
Copilot currently keeps in context the file you are editing. Cross file support is coming but not here.( https://githubnext.com/projects/copilot-view/ ). But it would be very very useful. One logical concept that&#x
333.
by itissid 3y ago
One thing that is surprising is just how vast is the chasm on Google/Bing Search quality for developer resources(and probably many engineering field with vast user documentation) and how good ChatGPT-4 has become on it. As a developer
334.
by itissid 3y ago
Congrats to all of the team at ISRO. Assuming the software development for these is non trivial; Its interesting that with the last mission(per Scott Manly) was designed to always select a spot(or more precisely a trajectory) and try hit it
335.
Tell HN: Have stopped using Search Engines to undrstand non-trivial prog. topics
1 points by itissid 3y ago|3 comments
336.
by itissid 3y ago
Its only if the cost for a FAANG stack is high. Cost of tech is always decreasing. If you think you are going to have a FAANG stack protected by crypto hardware attestation you will be beaten by the next guy who never needs to do it.
337.
by itissid 3y ago
Companies that want Physical Humans to watch what is on their websites(ads+content) and wall off said content at the same time. Every one of these companies is on borrowed time. Eventually AI will be able to read frame buffer and simulate m
338.
by itissid 3y ago
TL;DR: Chat GPT summary of 5 "pages" of the thing: https://chat.openai.com/share/bd6eac38-5bac-4c6f-b405-7ca7d8...
339.
by itissid 3y ago
Overheard at the watercooler: The only way a CS guy knows how to prove an algorithm works/does not work is to find counterexamples.
340.
by itissid 3y ago
I have noob thought on the potential of these in Formal path planning. Specifically given a set of functions that basically map {State -> Actions} given preconditions, transition functions (heavily paraphrasing STRIPS[1]) can a correct
341.
by itissid 3y ago
I write my thoughts in table and for each abstraction/concept in a cell and ask myself if I need to understand a sub concept to implement it? And put it in a cell in the next column one row below. Sorta like: https://pastebo
342.
by itissid 3y ago
Not to be coy or disrespectful( and this is excellent work and a way to learn new things) but if one is generating a parser as a means to an end(the end being execution of the AST on the database), what are the learnings one gets from just
343.
by itissid 3y ago
The cost to implement a model that can do P(Y|X) can unsurprisingly be much lower than something that can model P(X, Y) and that's not just because its an easier problem but economics and what just works easily. But in the world of the
344.
by itissid 3y ago
At its core, How closely related is this to Hamiltonian Chain Monte Carlo(HCMC) methods? Reading about waiting to reach equilibrium state, extracting paths and integrating it(IIRC this is analogus to estimating the expected value) in a spac
345.
by itissid 3y ago
One thing people do/did with search engines was to observe their ranking stability over time on common queries like "Cats" or "Dogs". If there was any change one could meaningfully investigate what has changed. Thou
346.
by itissid 3y ago
What is 100x launching on August 15?
347.
by itissid 3y ago
> My approach involves learning what's necessary to solve immediate issues and then moving on. When you move on do you crystallize your learnings like in a document or a chat with other people? Does it give you an idea of where to g
348.
by itissid 3y ago
City wide Town halls where every one can listen in but pre-registered people can ask questions are a productive usecase for public information. Those buildings can't accommodate 500 people.
349.
by itissid 3y ago
A Question. NYC has a law regarding surprise medical bills:( https://ag.ny.gov/resources/individuals/health-care-insuranc... ). If I understand that law correctly hospitals/facilities can't bill both in a
350.
by itissid 3y ago
> Groups with more negotiating power, like large hospital systems, are able to command higher rates than private practice physicians. A thought: I wonder if this also mean that large hospitals will have doctors that earn more and somehow
351.
by itissid 3y ago
Noob Thought on why OpenAI collaborated with MS: 1. I have found ChatGPT is very good at parroting back decent quality from learning from structured data. For example, learning reasonable code completions as well as answering technical ques
352.
by itissid 3y ago
Like to clarify, a new grad students could look at the current group and ask "Hey I know you are working on LLMs, but how many $$ of your grant are dedicated to how many TPU hours per grad student?"
353.
by itissid 3y ago
Noob Thought: So this would be a blue print on how a mid tier universities with older large compute cluster ops could do things in 2023 to support large LLM research? Perhaps its also a way for freshly applying grad students to look at a un
354.
by itissid 3y ago
That arxiv link to the paper is also the same guy I mentioned in my twitter reference. Subbarao :)
355.
by itissid 3y ago
An immediately useful application of this is in roombas that can not just clean the floor but effectively avoid/move obstacles. All the vaccums i have gotten generally get stuck or suck in paper/usb cables and get stuck in corners
356.
by itissid 3y ago
It would be interesting to see how it solve AI Planning Problems like BlocksWorld[1]. I've read that in the past with these things[2][3] when multiple goals needed to be met at once and there was interaction between them, it just falls
357.
by itissid 3y ago
I am a bit confused. This sounds a bit stupid. If this is DRM for web content, then what would non chrome browsers like FF for web pages that ask for such authentication? Refusing to work with other browsers than chrome seems like a lawsuit
358.
by itissid 3y ago
Fails to start the Sussman anomaly. https://twitter.com/sg3487/status/1681374390448009216?s=20
359.
by itissid 3y ago
I have pretty bad ADHD, back in the day when I was flipping between editors, I had to carefully catalogue what I want from an IDE and then force myself from ignoring everything else. Here is what my list chiefly has: 0. Reasonable Vim key b
360.
by itissid 3y ago
One important thing to not do here at any cost is to make a comparison between GPTs and KNNs and go "GPTs or BERT are meh like Gzip". Remember that GPTs, for all their flaws[1], are learning the joint distribution P(X,Y) which can

More ›