urlwolf

I had to stop biking to work (40 min each way) because I realized that I will have an accident if I continue, and I cannot afford the downtime.

How come this is not bigger news?

Argh, it dropped one feature that I loved: clicking on current tab does a ctrl tab.

You can read this the following way. The 'analogs', represented by Trump, realize they are losing power, that regulation may be their best shot even if that has proven to not work well, and are launching a full-force attack against the 'digitals' (eg Valley).

This event is more meaningful for the world economy and safety than North Korea playing with nuclear weapons. The sides of a possible WW III are analogs vs digitals.

What do you think? I'm thinking about writing a blog post expanding on the idea.

ML that helps me learn faster as a human. Let's be honest, watching videos online is suboptimal. The middle-ages method of listening to a professor in a large room is suboptimal.

Can you personalize what I learn? Can you find the exact best explanation for my current level by looking at my facial expression?

Well, hyperparameter optimization is pretty much that. There are people using algos to improve 'creative' tasks like circuit design.

Otter browser is pretty fantastic. Use the weekly for the best results.

They seem to have removed the 'audit course' option, at least on my setup. They may be doing AB testing.

Hmm, doesn't seem to log into slack. Linux version, FF as default browser. It takes me into the web app of slack but eul stays empty.

Agreed. I did my research and SG is not sustainable for my vertical (data science)

I asked support and they said no. I'm leaving quip.

Ardour if fantastic. Its mixer is better and more usable than pretty much any other DAW I've tried. They are competing with software that costs hundreds of dollars and winning in many aspects. They should be an example.

Sounds intriging. Any chance it would work under linux?

Scala far more than Clojure at this point.

Python has good bindings to venerable libraries in C and fortran that nobody would want to implement from scratch. Python libs like numpy and scipy were very convenient ways to call the highly optimized code in say BLAS. Then scikit-learn came and made machine learning easy for the masses. Before, libs in R would provide different methods but they were anything but unified (things like CARET notwithstanding).

In the last 2-3 years the main driver was deep learning. Since the whole point was to run on a GPU and CUDA remains a pain of low-level C, python made it very palatable.

R has lost the race by not jumping on the deep learning train. Scala shows promise because of Spark, its killer app.

The article is in German, but Google translate does a good job. You only need to read a few paragraphs to get the point. The fact that this newspaper (which usually exalts the virtues of anything made in Germany) has published this paper sets a new landmark. There's a 'before' and 'after'. In Jan 2017, Germans publicly admitted their car manufacturing industry is in trouble.

I agree wih the rest : this is too hard. Some papers will take a long time to grasp. More so after we have been conditioned to read things fast on he web.

Are there any deep learning libs that work ok AMD cards? Looks like CUDA totally owns that market.

To anyone with these problems: try Solus. It's far better than ubuntu (benchmarked), has almost no papercuts, and it runs games quite ok.

I would love to know the answer to this too.

Codekitt | Berlin, Germany | Onsite, but remote possible

- Senior PHP developer

- Junior PHP developer

- Senior frontend

- Junior PHP developer

Codekitt wants to enter the 'disrupting recruiting' market.

If you are a dev, think about your friends: Who has gotten killer training from a company they were working for? Noone? There's a reason for it. Because we switch jobs every two years, for a company it's very hard to justify spending money on training people. But they do want everyone to be on top of the latest and greatest in tech. And we do.

Because we devs like learning, we spend our evenings and weekends learning new technologies. Something the company we work for benefits from (that is, they get the knowledge on the lastest tech without having to spend money on training).

This is unfair. But this is how the world works.

Companies don't have budgets for training (or have say 500eur/person/year)... this is scary because the fast-changing tech world will leave you behind unless you spend time improving yourself. If the company doesn't spend money on training, then you have to find ways to catch up, probably over weekends and evenings. But companies do have big budgets for recruiting. Everyone assumes that you have to spend money to get good guys, and recruiters charge a good chunk of the first year salary for the hire. Companies budget for this.

We want to get around the fact that there's no budget for training, by taking it from where businesses do have a budget: hiring. So it's mostly budget relocation. But there's a clear benefit for the company: they get training 'for free.' And they get onboarding partially solved. It's common to hire someone, and the assume that the first month is wasted, with the person getting used to the tech stack the company uses. We remove a big chunk of the problem.

_No time for bullshit_

- We pay market rate salaries

- We care about you being fast learner, not '10 years experience on technology X'

- We pass the Joel test

- You will work on pretty mundane stuff at first (come on, you know a lot of web deb is mundane), but sooner or later you will start integrating machine learning APIs that we build in-house

- If you interview, we will not waste your time

You can apply here: https://dsr.recruitee.com/o/backend-full-stack-developer

Asus Q534UX looks pretty tempting. Metal body, nice specs. Unknown how well it'd do witn linux though.

I have a soon-to-be-4. Is touchscreen a must? I'm not sure he will be very good with mice or trackpads.

Probably I didn't get this, too tired, but would any product trying to take over the system's package manager be destined to fail?

Great topic. I have had this problem may times over.

Nope, sorry. This was an informal conversation with a German friend. He probably doesn't know for sure what patent this is. It's not something that is on the news.

Bosch, a German company that serves the auto manufacturers, seems to have bought patents on a battery technology that could be a breakthrough. Remember that 'the German way' of doing business is not to make noise at all. There have been entire German companies operating without marketing and sales departments, and there still are.

I live in Germany.

If anything, this analysis under-represents the success of solo founders. Because VCs don't like solo founders, it's less likely they go for a VC path. There must be plenty of successes that are bootstrapped. And these solo founders may not even bother adding their company to Crunchbase.

I'm one of those.

I have been in Berlin for 8 years and would not change it for London at all. In data science at least there's far more action here. More interesting technologies (Flink), people taking more technological risk and reaping the benefits.

10 years in academia, doing modeling/machine learning. 3 years doing moonshot startups. 1 year consulting in data science. 3 years doing a business that actually works, generates lots of value to many people (Data Science Retreat).

1- I would have gone sooner with a business that targets an audience I'm part of (HT to Amy Hoy for that). It's hard to fail that way.

2- I would have spend more time thinking about how to have fun. It doesn't happen by accident. We spend so much time designing our business and so little designing our life.

3- I'd have spent more time solving real problems.