mobileexpert

Stock appears to be soaring after hours.

I don’t really understand this, doesn’t YC strongly encourage investors, to the point of punishing those who violate, to wait until Demo Day? That way they get to build good market conditions for raising.

The part about Google not respecting your time is I expect pretty ubiquitous among big tech cos. The hand off between outsourced sourcing (LinkedIn searchers) to inside co recruiters is often a fumble point, so is the time between passing a screening and next step in the interview process. Whenever it seems you move between one functional area and the next (sourcing,recruiter,interview, packet/candidate review) there is some unnecessary (from the candidate perspective) waiting. I think this is unlikely to ever get better.

This thinly veiled advertisement gives me almost no confidence that the posting firm has ever met a data scientist much less possesses expertise in how to effectively recruit and build a function for them at a small tech company.

The first things to understand is if you are hiring for a DS role to support product development/understanding, growth or marketing, advanced BI, core product features (eg ML in your offering) etc. What VP your DS would report to in a 500 person version of your org chart makes a huuuge difference in understanding what skills and aptitudes you are looking for.

Design the position and responsibilities now and in 2 yrs for the ideal candidate, then design your hiring strategy.

Found one on Twitter: https://twitter.com/dercentralist/status/1213216360630759431...

Looks like it was the vendor ‘Legalinc’.

Strange to not see an official statement and post Mortem from Stripe mentioned anywhere. Can someone who got a letter post a (redacted as necessary) scan of it?

Looking at the Azure networking numbers I wonder if they got their configs correctly optimized. Azure networking was always tricky to get best performance when I used it heavily at my last co.

One of the aspects related to an unacknowledged managerial role with chores, is that I find it often comes with caring about the outcome more strongly or more specifically than others. Expecting everyone to care about your hobby horse as much as you do is a recipe for feeling overburden and under-appreciated. If people are happy with your work NOT happening then your “emotional” or other labor will not be something they find valuable.

See also https://en.m.wikipedia.org/wiki/Jensen%27s_inequality

You can see some related but not directly comparable efforts in HuDI see https://eng.uber.com/uber-big-data-platform/

Or https://iceberg.apache.org/

Which both keep track of data versioning and management of file based datasets on object stores.

Most people today have very ad how approaches to handling data versioning and lineage on Hadoop datasets.

I don’t think Amundsen is comparable. Amundsen is essentially a data catalog that has ambition to get into master data management.

Delta here is adding features more closely associated with RDBMS or MPP data warehouses to the Spark data pipelines with parquet data on object stores big data world.

The Delta here isn’t so related to CDC. This is a replacement for the ad how methods people use now to keep multiple readers and writers of datasets in parquet format in agreement when using Spark’s dataframe API.

Other efforts in improving the parquet datasets on cloud storage world:

https://github.com/apache/incubator-iceberg

https://github.com/apache/incubator-hudi

Happy to see Delta go open source.

I read more into this being a Thanksgiving news dump than any particularly scandalous details.

Facebook is embarrassed by what many orgs would consider routine corporate communications practices. This is positive that they care about their brand and reputation at a higher level than your local closely held global conglomerate.

A refrain that is very true in NYC.

It is a bit off-putting (and just wrong) to read the author talk about white male privilege being a comparable head start to her half a million dollars of family money she’s received. (And 28,000/yr still coming!) For an essay centered on owning up to personal privilege it still has some serious short comings in recognizing how extreme her wealth privilege is versus almost all intrinsic characteristic in society privileges.

This motivated me to reread the commons clause again. I remain so disappointed in the naming and positioning of it. If you want to sell commercial modules that is totaly fine, but Apache 2 with commons doesn’t jump off the page as an unfree license (which prima facia it is). Should be called the ‘free for personal use’ email us for pricing clause instead.

I don’t know the exact specifics, but the notion of co-founder vesting schedules exist today. This provides a simpler pre-agreed upon way to handle that kind of situation. If I recall correctly Micro-Soft wasn’t even a properly incorporated entity for the DOS deal making things like a founder needing to stop having an execution role for health reasons hard to deal with.

I assume Azure and AWS have some buddying up with Intel/Nervana and Nvidia counterstroke to Google TPUs. I can’t quite imagine what it will be though.

It takes around 20 minutes to spin up HDInsight clusters. So it is not instant but I wonder if databricks on Azure can do any better given they are running in customer environments so wont have a big pool of preallocated multitenant machines available.

It would help if you scoped your request to the parts of the stack you are looking at help with.

For instance if you are interested in soup to nuts sensors in the field to predictive maintenance software in the cloud, is very different set of companies than if you are just interested in data collection layer, signal transmission layer, storage layer, business logic or analytics layers.

You will find start ups that claim to be "full stack" but you might want to have some caution and understand what they've actually built and where it is in "in production" not in pilot or PoC phases.

The other interesting thing about IIoT start ups is that many are focused on specific verticals/application areas.

Companies involved with tracking trucking fleets versus operating drone fleets versus monitoring oil fields seem unlikely to all be relevant to you.

I want to see much better human in the loop systems in general. A simple example is systems that will allow users to tweak the output of creative/generative systems. E.g. Neural style transfer methods that allow the user to easily adjust the output with human centric feedback. I want to say "good start, but make the background more like Caravaggio would than it is now, and lighten up the foreground to be more playful" and have the algorithm adjust the output accordingly.

We aren't impossibly far off from this, but not that close yet either.

This uses the pushshift.io dataset. The dataset is collected by an independent non-Reddit inc user @jasonbaumgartne on twitter. Huge props to him for collecting it. Anti props to the authors for not giving him a better acknowledgement in the paper.

Azure ML

"The JIT (‘Just In Time’) byte-code compiler is now enabled by default at its level 3. This means functions will be compiled on first or second use and top-level loops will be compiled and then run. (Thanks to Tomas Kalibera for extensive work to make this possible.)"

Very exciting.

Just bought his Free Speech for Me but not Thee. Hadn't been aware of his work but this obit made it seem worth a read.

I don't know the facts in great detail on the surface it seems like a good verdict. However the pre verdict punishment of being banned from the Internet for three years while the case progresses is ridiculous. This is a great travesty to any sense of justice or fairness in judicial proceedings. The crown should not get to impose such conditions.

What about Twitter's none-core products? GNIP and Fabric (crashlytics and Answers)

NVidia should also market this for people who want to do molecular dynamics and other gpu enabled physics sim locally.

There is no way this is calculated on a 'fair' basis. There are around 3 million apps total between different platforms for mobile. This 1M number must be calculated counting versions of apps as new apps.