BenoitP

12 ms·

91.
by BenoitP 2y ago
OP is also the author of the popular dimensionality reduction algorithm UMAP. I guess the pipeline was embedding documents with an LLM (or even plain old word2vec average over the abstract might do it), and then reducing that to 2 dimension
92.
by BenoitP 2y ago
Pretty interesting! A big surprise for me was to find the explainability cluster quite far from the causality one. But I guess it stems from a cultural facet: causality is mainly the purview of statistics (with Pearl at the helm) with a str
93.
by BenoitP 2y ago
What a wonderfully crafted piece. But a part of me can't refrain from saying it would have looked way more elegant and succinct in 3D Projective Geometric Algebra. Most of the last sections (all intersections) feel like corner cases, w
94.
by BenoitP 2y ago
That'd be great if it could absorb the surplus, but right now at 12GW production it is curtailed at 0.29 GW. This means neither pumped storage, nor batteries, nor exports can deal with it. On one hand it might create a new market, hope
95.
by BenoitP 2y ago
I did something akin to that with the following monstrosity of DevOps: - set up Kafka - set up Flink - CDC on your DB, into Kafka - have your server maintain a websocket to clients, read Kafka from it, and dispatch messages according to a f
96.
by BenoitP 2y ago
It describes all the way the SQL could be executed, then choses the faster plan. For example: if you're looking for the user row with user_id xx, do you read the full table then filter it (you have to look at all the rows)? Or do you u
97.
by BenoitP 2y ago
According to the framework the author uses, how would you describe these in terms of state space management (apart from immutability, which the author dealt with)?
98.
Stacksort (2013)(gkoberger.github.io)
2 points by BenoitP 2y ago|1 comments
99.
by BenoitP 2y ago
This semi-old internet artifact, to be viewed in light of today's LLM training data. Once again XKCD was not much off the mark.
100.
by BenoitP 2y ago
I did a similar transition as a freelance, it was not a simple task. I started 7 years ago. First, I got lucky to be doing data close to where a business case allowed it. Then had to fight my client so that I could _improve_ his best sellin
101.
by BenoitP 2y ago
3Blue1Brown is really a godsend. I had been trying and failed to get an intuition for the self-attention mechanisms, and also why you want to stack you transformer layers. I went through a lot of material (Key-Query-Value interpretations, S
102.
by BenoitP 2y ago
Elon probably said on both occurrences that robotaxis and the 25k$ cars were priority #1. Anyway it doesn't hurt -even if this is a lie- to say the cheap cars are coming. It sucks financial derisking off of competitors. To competition,
103.
by BenoitP 3y ago
Yup. This is more of a signal on 1) experience with new EUV processes and 2) cost of labor. That's not taking into account the expensive EUV amortization. A hard truth I've heard in CEO circles is that western middle classes are f
104.
by BenoitP 3y ago
Hi, this is Elon Musk himself. Please refrain from outing publicly my relaxing bug-fixing hobby. Thank you.
105.
by BenoitP 3y ago
SEEKING WORK | Paris, France | Remote-open --------------------------- Machine learning engineer, specialized in Explainable AI / ML. 13 years professional coding experience, 24 years personal. Recent highlights: * Seamless integration
106.
by BenoitP 3y ago
Machine learning engineer, specialized in Explainable AI / ML. 13 years professional coding experience, 24 years personal. Highlights: * Seamless integration of hand-crafted rules with ML-build rules * Intuitive, visual data/signa
107.
by BenoitP 3y ago
Damn, Stonebreaker, Zaharia. Seems like the big guns are out for the new data model in town. This tech feels like fancy RDBMS triggers backed (and made unique) by a distributed log. Looks very much like the same tech as Flink's statefu
108.
by BenoitP 3y ago
I guess some risk analysis can be made with the plane's and Starship's cross-section. Let say they are cubes of 30m each. The expected area where they might both be present to be a 5km square. That's a 0.000025 chance of coll
109.
by BenoitP 3y ago
FR24 shows a Dassault Falcon 900EX doing circles around the expected Starship re-entry location (Perth to Perth flight plan): https://www.flightradar24.com/MXJ/345b8f09
110.
by BenoitP 3y ago
> 26000 km/h > 27,900 kph Seems like they want to test the limit. Same speeds as LEO, but guaranteed to come down.
111.
by BenoitP 3y ago
It is, how to reproduce: Twitter stream page, F12, network tab, look for m3u8 file, right click, copy url, open in VLC
112.
by BenoitP 3y ago
Updated, Thanks!
113.
by BenoitP 3y ago
When: 8:25 AM CT Launch window: 7:00 AM CT - 8:50 AM CT --- Updates: (future)T+40: Starship relight and entry T+12: Elevator music engaged, please stay tuned for T+40 T+11: Payload door testing T+8: Upper stage SECO, nominal orbit insertion
114.
Starship's Third Flight Test [video](spacex.com)
499 points by BenoitP 3y ago|395 comments
115.
by BenoitP 3y ago
Well the comment by pnin made it about GR, when the parent by boringuser2 was about the general contributions to relativity, both SR and GR. I believe "general theory of relativity" to be different than "general relativity&qu
116.
by BenoitP 3y ago
Which is the source for General Relativity, which suffered too from an attribution dispute with Hilbert's works. My point being: science doesn't happen in a vacuum, and is not totally ordered
117.
by BenoitP 3y ago
> Neither Poincare nor Lorentz are relevant to the genesis of General Relativity Well, that's just plain wrong. From the horse's mouth: > As we know, this is connected with the relativity of the concepts of "simultaneit
118.
by BenoitP 3y ago
Relativity was ripe for discovery. And a lot of other scientists came very close: https://en.wikipedia.org/wiki/Relativity_priority_dispute But that's just how science work, and no ill will would have been employe
119.
by BenoitP 3y ago
Plagiarized is too strong of a word. Poincaré too based his work on Lorentz's. And both with Einstein he sort of derived E = m c * 2, independently and earlier. But Einstein's publication was more complete. Science is not totally
120.
by BenoitP 3y ago
> It's absurd to tax something that can so easily migrate to a different country It creates an incentive in that country to tax it. If you have followed EU's Carbon Border Adjustment Mechanism, it is exactly how the EU intends

More ›