olavgg

31 ms·

1. 331.▲
   by olavgg 9y ago

   Since Unity became the default in Ubuntu, I've been using KDE. Was very happy with the 4.x, but not so happy with the bump to 5.x. I've been using 5.x daily since the beginning. Mostly it works really well, though I still have to
2. 332.▲
   by olavgg 9y ago

   This is so true, in fact you may need to rely on several cloud providers. For example we do question classification with our own model, but rely on Googles Speech Api(norwegian language) for speech to text so we can classify the question an
3. 333.▲
   by olavgg 9y ago

   We already run a polyglot stack at our company, and we use Docker(nvidia-docker) for our Python environment. With Java there is no need, and it is a lot less work updating and upgrading the JVM and our Java applications. I would use Docker
4. 334.▲
   by olavgg 9y ago

   Honestly I don't feel there is any need to run Java in containers. The war/jar file is its own container with its own dependencies. The JVM still makes the same syscalls as it would inside a Docker/Kubernetes container. In fa
5. 335.▲
   by olavgg 9y ago

   With the ClickHouse OLAP db I'm limited by memory bandwidth.
6. 336.▲
   by olavgg 9y ago

   Isn't TCP packet shaping normally done on the network interface these days so you have CPU offloading?
7. 337.▲
   by olavgg 9y ago

   I'm a varnish user and I have no complaints. It is well documented, easy to use and solves complex problems with VLC/VMODs. I'm not pushing nowhere near 100GB/s, but I have benchmarked up to 10GB/s which isn't
8. 338.▲
   by olavgg 9y ago

   We are 10 developers just using vanilla Javascript here. No mess, and a joy to work with. We use good old design patterns like observer, proxy, composite and so on. We also use inheritance and template literals a lot. The key is good docume
9. 339.▲
   by olavgg 9y ago

   I shop a lot on Ebay, I think it really works well.
10. 340.▲
    by olavgg 9y ago

    I really like the idea of serverless, in my use case a serverless setup would be great to run prediction on. For example a chatbot where you predict the intent. It would be a lot more cost effective than running the model on their own GPU&#
11. 341.▲
    by olavgg 9y ago

    You can get Mellanox ConnectX-2 on Ebay from 12USD, these are excellent enterprise NIC's. They also consume little power as they use around 5 watt.
12. 342.▲
    by olavgg 9y ago

    I've stopped reading database benchmarks, because they are extremely vague. Instead I spend my time optimizing my current solution/stack. For example Postgresql has hundreds of knobs that you can adjust for almost every scenario y
13. 343.▲
    by olavgg 9y ago

    Thank for this suggestion, I have just read the introduction and first chapter. This is my current company in a nut shell!
14. 344.▲
    by olavgg 9y ago

    That book is awesome, should be read by every Java developer! Discussions and examples from this book is often on Stack Overflow too, where the discussions go deeper and may be clearer.
15. 345.▲
    by olavgg 9y ago

    https://www.servethehome.com
16. 346.▲
    by olavgg 9y ago

    I think Python with scikit-learn, nltk, numpy, jupyter and finally Keras for deep learning is a great combination. I highly recommend learning Keras before you dive into Tensorflow or PyTorch. Currently there are some excellent video tutori
17. 347.▲
    by olavgg 9y ago

    I'm from Norway and working for a startup. You do not have to pay tax for the value of your shares, but for the assets that the company own. For a tech startup that means laptops and workstations.
18. 348.▲
    by olavgg 9y ago

    I wholeheartedly agree with you. This is something which I've been struggling with the last 6 months telling a colleague. He always want to use things that are completely new and sometimes just weird . He always believe this new way o
19. 349.▲
    by olavgg 9y ago

    Boost Ai | Backend, Full-stack, Devops | Stavanger, Norway | Full time | ONSITE At Boost Ai we create chatbots for the enterprise market. We have created our own NLP model and we create the data for learning in-house. This give our chatbot
20. 350.▲
    by olavgg 9y ago

    I've done plenty of benchmarks with LZ4 and GZIP compression. It really helps a lot, especially LZ4 should always be enabled as it is always faster. GZIP is fantastic for some datasets where you do sequential scans of many tables aroun
21. 351.▲
    by olavgg 9y ago

    Write endurance for Optane 900p 280GB 5.11 PB Write endurance Intel S3700 200GB 10 drive writes per day for 5 years or 3.65PB for 5 years. Write endurance Intel S3500 240GB 140 TB So I cannot understand why you think the write endurance is
22. 352.▲
    Exploring ZFS ZIL with Intel Optane and NAND(servethehome.com)

    121 points by olavgg 9y ago|43 comments
23. 353.▲
    by olavgg 9y ago

    The big challenge is creating good data. We constantly analyze chat conversations. Sometimes our chatbot answer wrong because of a corner case that we didn't think about. It could be lack of training data, context mismatch or that foll
24. 354.▲
    by olavgg 9y ago

    I wouldn't say niche, there are hundreds of use cases. I believe we have just scratched the surface of possibilities yet. I think you could say, in every situation you would ask a human a question you could also ask a chatbot. For me s
25. 355.▲
    by olavgg 9y ago

    I know that many on HN really dislike chatbots because they don't understand their super special case. As a chatbot maker, with several different chatbots that are live, I can tell you that I've seen the data. Which concludes that
26. 356.▲
    by olavgg 9y ago

    The biggest reason going for an enterprise vendor is support! As long you pay them big money you will get something that just works without worries. If something should go wrong, they will be on your door ASAP and fix the problem for you. V
27. 357.▲
    by olavgg 9y ago

    I always increase the seq scan cost, but another thing that helped me more was updating the table statistics. In that way you can make the planner better aware of your indexes. In my case I increased STATISTICS to 5000 and the planner immed
28. 358.▲
    by olavgg 9y ago

    There is a Norwegian startup that has decreased the latency dramatically. https://remarkable.com/ https://youtu.be/zpUPpiV7gAo
29. 359.▲
    by olavgg 9y ago

    We (Nordic fintech company) took the other direction, because one huge client on our customer list helps a lot for further sales. It is still true when we try to enter a new market / country, selling is easier if we have a big name alr
30. 360.▲
    by olavgg 9y ago

    Spring Boot Grails Rails Are all excellent technologies. Grails has excellent form/model binding and validation with what they call command objects.