bkanber

26 ms·

241.
by bkanber 8y ago
For low touch sales cycle company like Buffer, marketing is sales.
242.
by bkanber 8y ago
We don't know yet. But, for instance, the Cambridge Analytica data is nearly a decade old, and we didn't know back then that it could be used to (potentially) hijack a Presidential election. It's not so much a matter of what
243.
by bkanber 8y ago
It found my correct location automatically.
244.
by bkanber 8y ago
I use Processing whenever I teach an intro to programming workshop or course. It's kind of the perfect language for beginners. First, it eliminates the need to navigate the programming ecosystem. Rather than having to choose python, Ru
245.
by bkanber 8y ago
Overkill for the hobbyist, but ANSYS products like Fluent and CFX are what the professionals use. Quite pricey though.
246.
by bkanber 8y ago
It's not the privacy part of the law that's hard for companies. Most companies do not abuse your data. The GDPR is very expensive. Every data processing activity needs to be auditable, every customer relationship needs a data proc
247.
by bkanber 8y ago
We're a small business and it cost us about $50k to become compliant, if you add up the legal fees and hours spent on technology. We are also forced to get a more expensive insurance policy.
248.
by bkanber 8y ago
Having just lead the GDPR effort at my company, I disagree. It's absolutely a burden. It slows down sales and makes customer relationships more difficult and expensive. Any custom integration we do needs to have a custom data processin
249.
by bkanber 8y ago
There's a big difference between being able to calculate the gradient of an activation function you just invented and apply it to a backpropagation algorithm or analyze it for convergence, and simply understanding what a gradient is
250.
by bkanber 8y ago
AI is a broad field. Much of it is applied these days. Even setting a very high bar for what constitutes "AI", I would concede that Google Assistant, Siri, Alexa, and Cortana are AI systems. How many jobs are there supporting thos
251.
by bkanber 8y ago
Eh, I find that the current adaptive brightness works very poorly. Sometimes slight shifts in my sitting or holding angle will cause the screen to dim.
252.
by bkanber 8y ago
Data Protection Officer. There are a few situations where you MUST assign a DPO (large company, or systemic monitoring or processing of data, or processing of protected data). If you're talking about a tiny side project with basic data
253.
by bkanber 8y ago
If you have an email address you can give users for privacy requests and a promised turnaround time (we will respond to all privacy messages in 7 days) you're OK.
254.
by bkanber 9y ago
I agree. I don't think GDPR would ever be passed in the US because it is way too heavy a regulation and does absolutely disproportionately impact startups and small businesses. By disproportionately impacting small businesses you stifl
255.
by bkanber 9y ago
Yes -- the headline of this post does not accurately convey how extreme this move was. It should have read "abruptly shut down".
256.
by bkanber 9y ago
Many of these endpoints were scheduled for retirement in July and December of this year. Developers were NOT notified in advance that they would be abruptly shut down today. I also believe this is a reaction to all the recent privacy issues
257.
by bkanber 9y ago
The important thing to note here is that these endpoints were scheduled for retirement at future dates; some in July 2018, some in Dec 2018. Today, with no prior announcement, these endpoints were abruptly retired. Instagram had previously
258.
by bkanber 9y ago
This was not a deprecation. They immediately and without prior announcement revoked access to many endpoints that had been scheduled for shutdown later this year and next year.
259.
by bkanber 9y ago
The problem is that they threw out the deprecation schedule that they originally published and immediately retired/shutdown many endpoints with no prior announcements. Not even Instagram Partners were notified. For those reasons, I abs
260.
by bkanber 9y ago
This comment is really inaccurate. Under GDPR you must have a legal basis for all data processing activities. In the case of maintaining a social profile, the legal basis is "explicit consent" pursuant to "serving the interes
261.
by bkanber 9y ago
Basically GDPR says that you can keep backed-up data but you must have your backup retention age and processes defined. That is, it's ok to say "we retain backups and logs which may contain your data for 30 days".
262.
by bkanber 9y ago
The NTSB has control of the investigation and the data for now. That should be standard procedure IMO.
263.
by bkanber 9y ago
If you're #50 in the PGA in terms of average ball speed [0] and drive a regulation ball (46g)[1] at 175 mph (78 m/s), that ball has an energy of 140 J. If it accelerated from 0 to 78 m/s in 0.5 ms [2], it accelerated at 156 k
264.
by bkanber 9y ago
If you service any users in the EU or process any personal information of EU citizens, you must comply with GDPR.
265.
by bkanber 9y ago
Image feature recognition is not quite solved but I feel it's very close. It's easier, obviously, if the problem domain is very specific. In the past, like when I started on ML, the best tip was to make sure to do some edge dete
266.
by bkanber 9y ago
Every problem is different, so the only advice I can give is: research research research! Do the hard work up-front; figure out how to describe your problem in a mathematical sense, and identify the right tools to use for the shape of you
267.
by bkanber 9y ago
> How did you figure out what features would be useful? There are various feature engineering and feature extraction techniques. Filter methods, wrapper methods, and embedded methods. Principle component analysis, autoencoding, variance
268.
by bkanber 9y ago
This is an NLP problem if all you're trying to do is extract nouns.
269.
by bkanber 9y ago
The hype is and isn't warranted. ML is a much broader field than just neural networks. The hype for ML, in general, I think is warranted. We hit an inflection point when AWS launched and scalable processing power became cheap. It becam
270.
by bkanber 9y ago
I teach ML and am currently writing my 2nd book on it. I always advocate learning the fundamentals. Machine learning is math, and neural networks in particular rely on linear algebra and vector calculus. (You can build a NN without using

More ›