freakynit

8 ms·

181.
by freakynit 3mo ago
Update: Added UMAP... this is so freaking cool.. thanks for suggesting. It's on the insights page: https://hn-buddies.stupidlabs.lol/insights under "Author Map" heading.
182.
by freakynit 3mo ago
These seem fun to explore. Will definitely check out. Thanks!
183.
by freakynit 3mo ago
Yea.. even if not wrong, it's definitely not correct. Couldn't think of any other, so, sticked to it for now.
184.
by freakynit 3mo ago
Thanks for the UMAP suggestion. Will add. Most of the authors are actually missing. Full processing would have yielded multi-trillion row dataset. I didn't rally have that kind of compute with me. I have even tried running the cross-jo
185.
by freakynit 3mo ago
Thank you. This has been in my mind for past 1 year. Wanted to do it using vector embedding similarity match, but due to costs and compute requirements, had to resort to keyword based. The data comes from daily-updated public BigQuery datas
186.
Show HN: Hacker-News Buddies
6 points by freakynit 3mo ago|10 comments
187.
Show HN: HTTPS://HN-Jobs.com(hn-jobs.com)
2 points by freakynit 3mo ago|0 comments
188.
by freakynit 3mo ago
We may not see it yet, but, this is a way to capture government and use that to suppress competition and any dissent. The moment the government owns equity, every regulatory decision becomes conflicted. The top 3-4 AI companies in whic the
189.
by freakynit 3mo ago
Might be weird, but, I have heard this "2012" reference many many times when things started to change (for worse), and then "2020" when it all started to go downhill with an ever-accelerating pace. Very consistently thes
190.
by freakynit 4mo ago
Meme context: https://x.com/BrunoCendon/status/1926883516890898637
191.
by freakynit 4mo ago
80B + INT4 + speculative (FP8 mode) => 72,188 tokens/s effective ..da fck!!
192.
by freakynit 4mo ago
Same.. although rare, but have observed twice till date. Some blog post I read few weeks back said that DSV4Flash in xHigh effort beats even the pro model in xHigh effort.
193.
by freakynit 4mo ago
Recently read about some upcoming GPU chip vendor that speeds up multiply-and-add (the core of [pretty much any neural network, including LLM's) oepration by transforming them to log scale where the multiplication changes to addition.
194.
by freakynit 4mo ago
Just tried. Bare minimal and does exactly what's needed and nothing else. Love it. Thanks..
195.
by freakynit 4mo ago
You could produce a completely fcked up model using AI, and also a completely awesome one using AI. So, the "overstating capabilities" part really have nothing to do with this discussion. As for the claim part around AI being used
196.
by freakynit 4mo ago
You do realize this statement about "overstating model capabilities" has nothing to do with AI assisted AI development, right?
197.
by freakynit 4mo ago
"Assisting", not "one-shotting", a better model.
198.
by freakynit 4mo ago
What part?
199.
by freakynit 4mo ago
Read Anthropic's own blog post released sometime around last week.
200.
by freakynit 4mo ago
Nothing new in that. Everyone pays to someone or the other to make their own product/life better. Most of the times these products do not compete, sometimes, they do. How many times you believe duckduckgo would have google'd stuff
201.
by freakynit 4mo ago
These frontier models are now good enough that they can assist heavily with new optimizations for future models, including, code them. Restricting their usage to a few companies takes away that advantage away from other companies, thereby,
202.
by freakynit 4mo ago
Simple vector search + keywords + bm25 + text match + RRF. We specifically avoided graph construction due to associated costs. Everything is in just one sqlite file. Works fine for up to a few million document chunks.
203.
by freakynit 4mo ago
"To confuse your enemy, you must first confuse yourself" - Some reddit guy
204.
by freakynit 4mo ago
Would it be comparatively easy to make neuromorphic chips instead of traditional chips? I believe probabilistic algorithms like those employed by LLM's must be more tolerable to working with defects as well..?
205.
by freakynit 4mo ago
Thanks.. just looked it up. Seems super interesting.
206.
by freakynit 4mo ago
Man... I had to chatgpt your comment just to understand. But I do now. Basically, unlike current chip manufacturing process where every stage is deterministic and precise, the soup-world, the chemistry, is not. And we do not have accurate e
207.
by freakynit 4mo ago
Bacteria do that at scale, far far bigger than all chips combined. All it takes is chemical soup and a few starter seed dna's.
208.
by freakynit 4mo ago
This may be extreme, or, completely stupid, but, why are we not using genetics to "grow" chips in a chemical soup yet? Similar to Verilog/VHDL, don't we have some similar language to express circuits using gene sequences
209.
by freakynit 4mo ago
Damn!! I just tried a few links from navbar... all loaded in an instant like they have already been cached, even though they were not (since I use firefox in full private mode, and haven't visited their site in like months before today
210.
Is Trump-Netanyahu Rift on the Cards?(trump-netanyahu-rift.pagey.site)
3 points by freakynit 4mo ago|1 comments

More ›