vonnik

60 ms·

601.
by vonnik 10y ago
We've worked with Kyle Mitchell. He's a smart guy.
602.
by vonnik 10y ago
This is one of the best roundups of PR advice I've seen. (andreasklinger's points are also spot on.) Couple things to add: 1) Remember that embargoes need a journalist's opt in. You can't give them the news and order the
603.
by vonnik 10y ago
No. It's just that startups often don't do it well, and most PR firms don't either, but they still take your money. YC's advice is: don't waste your money or time, and if you're going to do it, try to do it rig
604.
by vonnik 10y ago
If we're talking startups, then the securities law you're thinking about doesn't apply. These are private companies.
605.
by vonnik 10y ago
I lived in Europe for 14 years. I had several Roma friends, and I also had several family members (grandmother, father) robbed by Roma in the street. I don't see much benefit to letting those who want to leave Europe live in America. M
606.
by vonnik 10y ago
Sorry, that should have been WWII, you're right. Anyone who wishes to know more about France's lack of preparation for both wars should read the French historian Marc Bloch, who fought in both and died in the resistance. L'Et
607.
by vonnik 10y ago
AI is moving fast. Faster than it has in a long time. In research circles, "supervised learning", or data classification problems, are considered solved. Consider the magnitude of that statement: If we have sufficient labeled data
608.
by vonnik 10y ago
Tom Fuller is one of the best reporters at The New York Times. He covered the Arab Spring in North Africa. He covered the military junta in Myanmar, sneaking over the Thai border. His life work deserves a Pulitzer, and now we're lucky
609.
by vonnik 10y ago
To follow up on what @agibsonccc has been saying: Remote, open-source hiring is one of the best ways to match companies and candidates. I used to have to recruit for a closed-source company, and it was really hard to get good information -
610.
by vonnik 10y ago
This article and the study it cites miss the point. Cities exist to make it easy for everyone in them to meet or work with anyone else as cheaply as possible. They are markets for jobs, goods and services. When cities make it difficult to m
611.
by vonnik 10y ago
Vanilla deep learning is so far ahead of what most companies use for analytics, that very little research is needed to offer a product with significant advantages over what's in practice now. In fact, I'd say most ML/DL/
612.
by vonnik 10y ago
Cyberconnecting cyber-readers with the cybernews: https://twitter.com/CyberEveryword
613.
by vonnik 10y ago
I was disappointed in this article when it was first published in July. Whenever a newspaper columnist makes a plea to reset humanity's "moral compass", you know that the article is basically worthless, an exercise in futilit
614.
by vonnik 10y ago
I like to think of people as having chronological age as well as a [skill] age, with chronological age being largely irrelevant among adults. I'm 41 and my co-founder is 27. He had been programming Java for 7 years when I met him, and
615.
by vonnik 10y ago
According to Aristotle, persuasion has three parts - logos, pathos and ethos. Your logical argumentation is logos, a powerful tool but not the only way to persuade. Pathos is the appeal to feeling (avoid triggers), and ethos is the establis
616.
by vonnik 10y ago
the usual plug for wolfram...
617.
by vonnik 10y ago
For most of the world, being CEO means being CEO of an established company rather than a startup. And really, incorporating a company and making yourself CEO doesn't amount to much. There's a lot of work to do before the title mea
618.
by vonnik 10y ago
fwiw, deeplearning4j is a much easier install: http://deeplearning4j.org/quickstart
619.
by vonnik 10y ago
Bringing some form of feature introspection to deep neural networks will probably involve clever ways of visualizing the feature activations of unstructured data https://arxiv.org/abs/1603.02518
620.
by vonnik 10y ago
Integration of Neural Networks with Knowledge-Based Systems https://www.uni-marburg.de/fb12/datenbionik/pdf/pubs/1995/ul...
621.
by vonnik 10y ago
That's what I was thinking of, too. If any one wants the direct link to the arXiv paper, it's here: https://arxiv.org/abs/1602.04938 . One of the authors is Carlos Guestrin, one of the co-founders of Dato&#x2
622.
by vonnik 10y ago
The popularity of comments about hard work on this thread make me think that it's difficult for people to accept studies like this. And for two reasons: 1) We are born with very different innate abilities, which leads to great inequa
623.
Intel buying chipmaker Movidius to boost artificial-intelligence efforts(marketwatch.com)
2 points by vonnik 10y ago|0 comments
624.
by vonnik 10y ago
[Disclosure: I wrote this.] http://deeplearning4j.org/neuralnet-overview.html Also, this book is coming: https://www.amazon.com/Deep-Learning-Practitioners-Adam-Gibs...
625.
by vonnik 10y ago
Great article. For anyone who wants a book-length drama revolving around the same pitfalls, read Jerry Kaplan's great account of GO Computing, published in the mid 90s. Time is on their side, and so is the Koolaid. https://w
626.
by vonnik 10y ago
Couple more data points: Intel recently bought Nervana Systems, a startup focused on creating a neuromorphic ASIC for deep learning, which had been financed by DFJ. Nervana wrote a fast open-source library called neon for GPUs. I expect the
627.
by vonnik 10y ago
Small sample size, ambiguous term, self-reporting... There's not a lot of meat to chew on here. I would posit that wisdom is taking the appropriate situation to achieve one's goals, as well as having the right goals. Obviously t
628.
by vonnik 10y ago
Are you comparing 2015 DL4J to 2016 TensorFlow? Your opinion of our framework seems outdated. You're welcome to try us now -- that would actually be a fair comparison.
629.
by vonnik 10y ago
Spoiler: 95% of them are Udacity students without experience or budgets.
630.
by vonnik 10y ago
this is fantastic, and leo is supersmart.

More ›