1/ 👋 Excited to share what we’ve been building at https://t.co/GOQJ7LjQ2t + we are going to tweetstorm our progress every week!

Week 1 highlights: getting shortlisted for YC W2019🤞, acquiring a premium domain💰, meeting Substack's @hamishmckenzie and Stripe CEO @patrickc 🤩

Important paper from Google on large batch optimization. They do impressively careful experiments measuring # iterations needed to achieve target validation error at various batch sizes. The main "surprise" is the lack of surprises. [thread]

https://t.co/7QIx5CFdfJ

The paper is a good example of lots of elements of good experimental design. They validate their metric by showing lots of variants give consistent results. They tune hyperparamters separately for each condition, check that optimum isn't at the endpoints, and measure sensitivity.

They have separate experiments where the hold fixed # iterations and # epochs, which (as they explain) measure very different things. They avoid confounds, such as batch norm's artificial dependence between batch size and regularization strength.

When the experiments are done carefully enough, the results are remarkably consistent between different datasets and architectures. Qualitatively, MNIST behaves just like ImageNet.

Importantly, they don't find any evidence for a "sharp/flat optima" effect whereby better optimization leads to worse final results. They have a good discussion of experimental artifacts/confounds in past papers where such effects were reported.

Read through a forum thread on Chastity Mansion to see who is talking about the Edge-o-Matic and figure out what more to address regarding frequent questions, as well as scout out user feedback. Here's some thoughts you can carry back to the forums!

My website doesn't have any product reviews, but people scatter them everywhere else in forums, tweets, xtube vids, etc.

If you ARE a previous customer, and you like the thing, can you drop a review on my site to help others?

The price is steep, but that's the amount of coins I calculated to make these. I have several spreadsheets calculating every expense involved with this project.

The current cost sustains production and development of new tech, and is helping create jobs. We are 2 people now!!

It IS 3D printed, because injection molding is orders of magnitude more expensive for the initial tooling, and I am still making small revisions to the enclosures.

Injection molded plastics are a six thousand dollar commitment for fifteen cent parts. Great for mass mfg.

Also, if you didn't know, 3D printing technology allows for some shapes which would be difficult, expensive, or impossible on injection molding or milled parts!!

Undercuts, which I use in this design, require special lifters to mold that cost a tremendous fortune.

NEW: @SidneyPowell1 radio interview w/ @toddeherman "The new report has a truckload of evidence of foreign interference in the election in it. That's what people seem to want to deny, even though the FBI & CISA ... issued an advisory and alert back on Oct. 30th & updated on

2/ Nov. 3rd, so for them to deny that now is absolutely absurd. Plus, we have an affidavit from an expert cyber-warfare [individual] who even shows the diagram of all the attacks coming in. We know that the VPN ... the Dominion people left those open and unencrypted on the night

3/ of the election, so that anybody could get in. We know that the packets of information went to Serbia [note: over 100 Dominion employees scrubbed there LinkedIn profiles of any affiliation with the company after the election, many of whom were in Serbia], Liechtenstein,

4/ Spain, Venezuela, Hunan, China & Hong Kong. There was traffic all over into our election system that is supposed to be self-contained, not accessible by the internet at all. We have witnesses who document that it was connected to the internet. We have a study out of

5/ [the Dominion voting system in Antrim] MI, they know that it was connected to the internet. They erased their audit files & their adjudication files, where they dumped over 68% of their ballots into a bin that they call "adjudication" [that would've been thousands + thousands

MAHARAANI TARARANI: THE QUEEN OF MARATHAS

Rani Tarabai was the brave daughter-in-law of Chhatrapati Shivaji and one of the greatest medieval Monarchs India ever had.
She was born in 1675 to one of the bravest commander-in-chiefs of Shivaji's Army named Hambir Rao Mohite.

Tarabai was just 8yrs old when she was married to Shivaji's younger son, Rajaram. Mughals and Marathas were constantly at war with each other. Shivaji was crowned the ruler of the Independent Maratha Swarajya in 1674 which was established by him.

However,the great leader died in 1680&the Maratha Samrajya went into a complete&total chaos.Marathas saw some very tough times ahead. In 1689,when Raigad Fort fell to the Mughal Army of over 15000,Shivaji's first wife Saibai and his elder son Sambhaji were captured&put to death.

Sambhaji's wife and son Shahu were taken prisoners to the Mughal court. During the same battle, Shivaji's second son Rajaram(Tarabai's husband) , who was the heir to the throne now, managed to escape using a disguise. He made his way to the Gingee Fort in Tamil Nadu.

Mughal General Zulfikar Khan, however, followed them to Gingee laying siege to the Gingee Fort. From 1690-1698,This was the longest siege by a Mughal Army of any Fort which stands testimony to the military and diplomatic credentials of Rani Tarabai.