000webhost

Web hosting

Tuesday, March 20, 2018

Social Media Bot Coding Notes, Random Stuff, and More

- I obviously dabble in the marketing space form time to time. Of late, I've been working with social media bots a lot. Here are my notes...
- you need to be able to code. It's plain nonsensical if you are using social media bots to increase your online presence but you have to wait for updates to come via alternative sources. One frustrating thing is that people could have pulled different version of code from source repositories online which means that patches may not be compatible with one another. The following is an example of this... It's quicker to fix things yourself then to rely on others. My guess is that commercial operators suffer from the same issue during code alterations upstream but simply hide the fact that their bots are broken from time to time?
- if you've built anything which has to interact with anything else online you'll be aware of the moving target problem. API's can change without any warning and you have to update your code. My guess is that this is one of the reasons why many developers give up on maintaining their code. They just can't be bothered keeping up to date with changes. In the meantime, they're struggling to find ways to monetise their code
- code quality varies drastically across the board. Things to look at are error trapping, diagnostics, commenting, functionality, documentation, etc... The ad-hoc nature of a lot of code out there makes it hard to hunt down and figure out problems in more sophisticated bot software. Do not expect to just download code off the web and have it up and running instantly and perfectly
crypto subreddit
- you need to keep pace with relevant social media networks and how they work. Gaining reputation on one network versus another is different. Instagram relies on pictures, Twitter relies on 'tweets' of less then 140 characters, etc... Each one is different but the core theories remain the same in order to increase reputation and popularity. Content... Given each social network is somewhat unique it makes sense that there aren't always analogous capabilities across the various social networks...
http://www.visontay.com/
http://sm4lg.com/how-to-generate-leads-on-instagram-with-12-insanely-easy-tips/
https://alejandrocremades.com/she-built-a-100-million-fashion-brand-by-using-tech-to-drive-sales-growth/
https://2startups.net/2019/10/17/corporate-venture-capital-vs-vc-cvc-investors-booming/
facebook python interface auto like
https://github.com/arsho/autolike
http://programming-in-python.blogspot.com.au/2015/12/auto-like-and-auto-comment-facebook.html
http://facebook-sdk.readthedocs.io/en/latest/
https://github.com/streeter/autoliker
https://devbest.com/threads/python-using-python-facebook-api-noob-friendly.74400/
https://bitsdroid.com/facebook-bot/
http://apental.com/
https://www.quora.com/How-do-I-write-a-Python-script-to-comment-automatically-on-each-post-of-a-Facebook-page
http://www.newtechcj.com/python/auto-like-and-auto-comment-facebook-posts/
facebook follow and like bot
https://kingdomlikes.com/
https://www.fblikebot.net/en/download
https://www.quora.com/Where-can-I-find-a-Facebook-auto-like-bot
https://qz.com/476914/i-built-a-twitter-bot-that-entered-and-won-1000-online-contests-for-me/
https://bitsdroid.com/facebook-bot/
http://danielbussius.com/build-facebook-chat-bot-integrate-infusionsoft/
- beware farm type operations. They're much more difficult to distinguish against real traffic and there's no guarantee that they'll help actual sales. It may help with ranking/optimisation though
https://www.abc.net.au/news/2019-03-30/farming-for-likes-fake-social-media-engagement/10944078
- in the end I gave up on finding a bot for Reddit (sorry, but I can't share the code for this for 'commercial reasons'). You may have to do the same and build one from scratch. Note the terms and conditions though...
reddit python bot github
reddit bot github
Reddit bot
reddit python bot
https://www.reddit.com/dev/api
https://www.reddit.com/r/Python/comments/78uvdw/i_wrote_a_reddit_bot_in_python_a_few_weeks_back/
https://github.com/aydwi/explainxkcdbot
git clone https://github.com/aydwi/explainxkcdbot
https://praw.readthedocs.io/en/v3.6.0/pages/writing_a_bot.html
https://praw.readthedocs.io/en/v3.6.0/pages/useful_scripts.html
reddit python bot 2018
https://www.reddit.com/r/RequestABot/comments/85rmt6/need_help_setting_up_a_bot/
https://www.reddit.com/r/RequestABot/comments/3d3iss/a_comprehensive_guide_to_running_your_bot_that/
https://www.reddit.com/r/botwatch/comments/2261dv/bot_creators_do_any_of_you_host_your_bots_on/?utm_source=reddit&utm_medium=usertext&utm_name=learnpython&utm_content=t1_dkck50e
bash reddit bot
https://github.com/HugoNikanor/reddit-parenthesis-bot
http://www.reddit.com/api/login/
https://stackoverflow.com/questions/15014210/how-to-login-using-reddits-api
login reddit curl
https://www.reddit.com/r/redditdev/comments/3w38h5/user_required_when_using_curl/
https://www.reddit.com/wiki/api
praw simple reddit example
https://pythonprogramming.net/introduction-python-reddit-api-wrapper-praw-tutorial/
increase karma reddit
https://www.wikihow.com/Gain-Karma-on-Reddit
https://www.wikihow.tech/Get-Karma-Fast-on-Reddit
praw reddit example
http://praw.readthedocs.io/en/latest/getting_started/quick_start.html
https://progur.com/2016/09/how-to-create-reddit-bot-using-praw4.html
- note obvious online scams regarding stealing credentials. Watch carefully...
facebook chatbot free
https://chatbottutorial.com/messenger-bot-builder/
https://www.chatbots-builder.com/free-facebook-chatbot
https://snatchbot.me/#prices_box
- use every obvious trick. f4f, followforfollow, follow4follow hash tags on your posts. Use mass add, follow other lists, get intersect of two obvious f4f to get those likely to followback, target active users,  target mutual interests/groups, etc... Use SEO techniques on your website for 'easy wins' and 'easy traffic'...
free followers chrome facebook
http://www.multifollow.com/fb-follower/
adding more facebook followers bot
https://blog.bufferapp.com/first-1000-followers-twitter-facebook-social-media
https://www.quora.com/Where-can-I-find-a-Facebook-auto-like-bot
- there are heaps of different types of 'bots' out there. CLI based, browser add-on/extension based, web server based, cloud/SaaS based, etc... One of the ones that isn't so obvious is macro based. It circumvents the API issue entirely and can be more easily updated then ones based on code. The main caveat is that they aren't as flexible. Be prepared to build multiple layers of code/abstractions to get things done...
facebook messenger bot
https://developers.facebook.com/videos/f8-2016/introducing-bots-on-messenger/?_fb_noscript=1
https://manychat.com/
https://manychat.com/pricing
https://developers.facebook.com/docs/messenger-platform
https://developers.facebook.com/docs/messenger-platform/getting-started/quick-start
https://botlist.co/bots/filter?platform=13
https://blog.hootsuite.com/facebook-messenger-bots-guide/
https://chatbotsmagazine.com/the-5-best-facebook-messenger-bots-you-should-try-immediately-d40f1acef767
http://www.adweek.com/digital/alex-tarrand-lucktastic-guest-post-facebook-messenger-bots/
desktop automation with image recognition debian
autokey python ocr
python to exe
alternative to sikuli
instagram bot chrome
- giving 'a personality to your bot' may be important to product branding so test it out...
https://www.ibm.com/watson/services/tone-analyzer/
- get used to bypassing or finding workarounds to security systems. This includes caching, anti-Bot systems, etc...
bypass google bot check
getting around cloudflare
- get used to huge amounts of code breaking. Don't be scared off by acronyms (XML, HTML, XHTML, CSS, REST, API, SOAP, JS, JSON, GraphQL, etc...). At the end of the day, coding is coding... and most high level languages don't really differ by much. Most of the social media bots I've come across require a lot of code re-writes just to run stably (they won't run continuously or without major errors without someone standing by or code modifications), commenting is poor, etc... and are generally forks of a core piece of code somewhere upstream, Times of major code changes at social media companies can be frustrating. Expect code to break all over the place... Following example is of instabot.py in the last few days
# mode 3, likers_protocol.py
        #if self.media_by_user[self.current_index]["likes"]["count"] >= 10 and self.media_by_user[self.current_index]["likes"]["count"] < 100:
        if self.media_by_user[self.current_index]["edge_liked_by"]["count"] >= 10 and self.media_by_user[self.current_index]["edge_liked_by"]["count"] < 100:
# mode 0, userinfo.py 
        #id_user = all_data['user']['id']
        id_user = all_data['graphql']['user']['id']
There is an issue on line 42 in userinfo.py. Key 'user' cannot be found. #1017
instabot github issue
- spend enough time on Github, Bitbucket, etc... and you figure out that most coders out there fork from existing code sets. Ironically, I've often found this to be counterproductive and is probably the inversion of my work? I have more original/clean code then forked code?
- exploit growth trends whenever possible...
https://www.smh.com.au/business/small-business/shop-gives-away-coffees-as-cba-s-payment-terminals-fail-20180319-p4z52t.html
- watch out for multi-factor authentication and anti-bot systems. If an error comes up that doesn't really make sense test all the usual avenues: Can you still login as a normal user? Can you login from multiple locations? Is the problem isolated to one version of code? Are errors limited to when your bot is running with a particular set of parameters? etc...
- over time, you figure out what the 'commerical operators' of social media bots have done. They run their code all day, they've basically optimised the living daylights out of existing code out there and may possibly be using bots on top of bots. That's the only way that it makes economic sense for them to reach the numbers that they do for the cost levels that they do. Super basic modifications can lead to enormous performance increases which are less likely to draw suspicion from social media companies. Active user detection: minimum number of posts, follower to followers ratio to increase chances of a 'follow back', pure speed, counter measures to ensure your bot isn't banned or shadow-banned, etc... Ironically, I've found that related accounts (subject wise) are more likely to follow one another then brute force followforfollow type scams. It is genuinely realistic for you to hit 'commercial operator' (low to mid level) type figures by simply optimising existing code from currently popular social media bots. To hit the high level operators type numbers it will take genuine time, work, and effort though...
auto like twitter
likeforfollow posts tweet
auto follow back accounts twitter
guarantee followback twitter
guarantee likeforfollow twitter
gainparty
gaintricks
- sound obvious but you need to test out the machine learning and artificial intelligence systems for chatbots before deploying them. Very awkward content at times otherwise (doesn't necessarily sound robotic but it definitely doesn't 'make sense'. It just sounds really 'strange'). Ideal situation is of course to start from scratch or build a custom solution though. Obviously, if you've dealt with this area you'll know that many 'core theories' aren't that much different from one another
- from time to time a lot of large scale 'bot operators' are get shutdown. If you examine their pricing models and contract terms it's obvious that many of them are worried about bankruptcy or future financial viability of their businesses. Feels like social media firms have continual and regular tests to figure out how much bot traffic there is?
Json Decoder #1153
https://github.com/instabot-py/instabot.py/issues/1153
massplanner
https://www.blackhatworld.com/seo/followliker-only-gets-me-20-followers-per-day-on-instagram.852508/
- a lot of automated marketing solutions out there now have automated social media marketing components to them as well. Obviously, many are shutting down operations as well though. One interesting theme is the general clamp down on pure commercial fare. It effectively means you need to mix 'normal content' in with 'commercial content'
- watch for traffic bot type scams... You're basically paying people for automated traffic to your website
traffic bot github
https://github.com/ecapuano/web-traffic-generator
- don't forget about copying over relevant databases if something goes wrong and you need to start over. This will stop you from having to cover old territory and reduces error rates as your software/bot contacts the social media company's servers which reduces the chances of a soft or hard ban. Most database backends rely on flat files or else SQLite (I've been intending to build a master-master replication systems for SQLite for a while now (for fun really) but have been bogged down with other work. May do this in the near future if have time)
- I've worked in the SEO, marketing, sales, search engine, online advertising space before and I'm seeing the same rough patterns. Basically, linear growth is expected. For a given amount of traffic the percentage level of engagement is roughly the same (think back to algebra and y=mx+c where: x=engagement rate, c=base traffic, m=average traffic and y=a measure of overall efficiency. That's what you're mostly seeing here. Engagement rates oscillate at roughly the 0.5-2.5% level. Really good campaigns will hit closer to 5% mark. Yes, I'm aware that a lot of firms designate 3% mark as being an indicator/marker of possible fraud)... If you genuinely want to have faster growth you need decent content in combination with bots, etc... to make things better. This is the main reason how/why individuals and organisations seem to have such a huge online presence versus others. They've just optimised things across the board
- software such as hootsuite.com or dlvr.it can help a lot to centrally manage things and also provide semi-bot type functionality
twitter hootsuite add followers
sudo pip install fake_useragent
instabot.py scam
instagram bot traffic
90% of traffic is coming from bots, how to get a refund? | Facebook
free alternative hootsuite
posting to instagram github
crypto rss feed
ico rss feed
twitter auto follow
- be prepared to deal with a lot of reconfiguration issues
praw subreddit configuration
post twitter bot github
using reddit-twitter-bot
"You provided the name of a praw.ini configuration which does not exist." #12
sudo pip install praw==3.6.0
- there is obviously a growing place for automated marketing type solutions out there...
- some things are easier to 'bot/automate' then others... Obviously, the easiest way to 'bot' youtube views is a random JavaScript type script (sorry, but I can't share the code for this for 'commercial reasons'). Remember to use proxies, throwaway emails, implement countermeasures to overcome anti-bot systems (such as Cloudfare, Re-Captcha, Captcha, etc...)... Note, some anti-bot measures are super simple to bypass. For instance, just switching browsers, proxies, user agents, etc... will overcome a lot of them. Protocol level anti-bot measures seem to be rare...
randomly open youtube videos from list javascript
https://stackoverflow.com/questions/22128689/javascript-random-youtube-video-autoplay
https://stackoverflow.com/questions/41624880/making-a-youtube-video-player-that-plays-videos-randomly-from-a-list-server-side
https://ccm.net/faq/918-javascript-choosing-a-link-at-random
increase youtube view count curl
increase youtube hits automated linux
youtube view increase automated linux site:github.com
youtube view increase linux site:github.com
youtube view bot
youtube view bot site:github.com
boost youtube views via curl
https://www.blackhatworld.com/seo/youtube-views-booster-bot.517864/
bypass cloudfare
https://blog.christophetd.fr/bypassing-cloudflare-using-internet-wide-scan-data/
cloudfare protected sites
1       upwork.com
2       coinmarketcap.com
3       codepen.io
4       bittrex.com
5       coinbase.com
6       medium.com
7       getbootstrap.com
8       news.ycombinator.com
9       fiverr.com
10      thepiratebay.org
https://www.wappalyzer.com/technologies/cloudflare
- core theories for 'botting' all social media services are basically the same whether we're dealing with professional networks, recreational networks, etc...
linkedin bot github.com
https://github.com/Kravalg/linkedin-bot
https://github.com/lambtron/linkedinbot
https://github.com/snmslavk/linkedin-bot
https://github.com/thepunkgeek/linkedin-bot
https://github.com/helloitsim/LInBot
https://github.com/cardangi/LinkedIn-BOT
linkedin bot site:github.com
linkedin bot python site:github.com
https://github.com/RyanOM/LinkedInViewBot
linkedin bot python site:github.com -selenium
https://github.com/thepunkgeek/linkedin-bot
https://github.com/kingdavis2324/LinkedInBot
https://github.com/toxtli/linkedin-invites-sender-bot
https://github.com/VarunAV/Linkedin-bot
https://github.com/fabricio-aguiar/Easy-Apply-bot
https://github.com/ozgur/python-linkedin
https://developer.linkedin.com/docs/oauth2
run selenium in background
https://stackoverflow.com/questions/16180428/can-selenium-webdriver-open-browser-windows-silently-in-background
https://www.blazemeter.com/blog/headless-execution-selenium-tests-jenkins
https://packages.debian.org/search?keywords=jenkins
managing company links linkedin
https://www.linkedin.com/help/linkedin/answer/1591/accessing-your-company-page-admin-center?lang=en
- graphics are a big problem if you aren't great with Photoshop or GIMP and your social media platform is heavily based on graphics (such as Instagram). Rely on automation to deal with this...
shattered glass effect imagemagick convert
Fred's ImageMagick Scripts
imagemagick scripts
cool free imagemagick scripts
- at the end of the day these companies are just technology companies...
https://au.news.yahoo.com/a/39561992/facebook-rocked-by-new-data-breach-scandal/
https://www.itwire.com/security/82197-uk-yet-to-obtain-warrant-to-raid-cambridge-analytica.html
https://www.itwire.com/security/82178-former-facebook-exec-says-data-harvesting-a-regular-occurrence.html
https://www.theguardian.com/uk-news/2018/mar/21/why-have-we-given-up-our-privacy-to-facebook-and-other-sites-so-willingly?CMP=share_btn_tw
http://www.abc.net.au/news/2018-03-22/facebook-cambridge-analytica-digital-surveillance-data-privacy/9575160
- ideal situation is that you can run them all day without having to worry about infrastructure concerns. Obvious candidates for bot hosting are AWS, Azure, Google Cloud, etc... You seem to need a credit card to sign up even though you may only be using the 'Free Tier' version of their services
https://www.infoworld.com/article/3179785/cloud-computing/aws-vs-azure-vs-google-cloud-which-free-tier-is-best.html
https://www.crn.com.au/news/nutanix-unveils-free-version-running-on-aws-google-public-cloud-412099
https://www.quora.com/How-can-I-register-for-Microsoft-Azure-without-a-credit-card
https://azure.microsoft.com/en-au/free/free-account-faq/
aws prepaid visa signup
https://www.reddit.com/r/sysadmin/comments/34h58p/aws_free_signup/
openstack free
https://platform9.com/blog/announcing-openstack-sandbox-platform9/
https://www.quora.com/Is-OpenStack-free
https://www.openstack.org/software/start
- bots aren't a replacement for decent content... Traffic only stays up as long as the bot keeps on running... What bots do out there is basically tap other users on the shoulder from time to time. You need genuinely viable content to get 'free traffic'...
- people think that having a targeted campaign is the easiest/best way to do things but you at the end of the day 'need traffic' to get engagement to get sales. There's no way you can get around this. Look at all of the social media accounts out there and do a comparison. They seem favour traffic first even if it comes at the cost of getting their brand up front and centre
most effective finance marketing campaigns
https://blog.bufferapp.com/marketing-campaigns-2016
most popular company social media accounts
- marketing people (particularly in your area of interest) will generally engage (likes, follow, retweet, etc...) more highly. Guerrilla marketing tactics will yield higher engagement levels as well...
- bot development proceeding smoothly. Certain accounts more likely to like, follow, retweet. Even more likely if areas of interest are similar. Marketers more likely to engage/retweet?
auto retweet twitter
#marketing
#lifehack
#growthhack
#instagood
- depending on needs you may need to combine code from multiple bots, start from scratch, extend existing bots, etc... Don't try to over complicate things. Building requisite bots isn't that hard. Core code is on my website/blog and via Github in other people's repos... If all else fails use Selenium (or equivalent product). Most of the time you'll be dealing with Python based bots. The rest of the time you'll be dealing with JavaScript, Java, and rarely Ruby. Use throwaway email addresses to ease account creation
auto comment twitter bot site:github.com
free increase twitter bot online
followforfollow site:twitter.com
followforfollow site:instagram.com
tweepy unfavorite hash tags
vickyqian/twitter crawler.txt
- there are lots of alternatives out there...
instagram bot
instagram followforfollow accounts
chrome extension instagram grow
instagram accounts which autofollow
marketing bot hashtags instagram
- there are often very distinct indicators of bot activity... High following to follower ratio, exceedingly low or high engagement rates, relatively low activity but high number of followers, etc... A lot of them seem to be propaganda and commercially orientated... Incidence of individuals using bots seem to be lower. Normally, lift is across the board if quality is relatively even. If not, it indicates that their may be quality control problems somewhere in the chain?
- there seems to be a lot sub-communities out there...
https://twitter.com/Cryptopiia_NZ
- if you have a 'standard media background' then you need to realise that social media marketing can be very different in spite of some core ideas/concepts remaining the same. One thing that is interesting for me is that my experience has taught me that people generally don't like being stuff that they don't like. It's obviously inverted when they are somewhat interested though...
twitter pool follower
https://foundr.com/how-to-get-more-twitter-followers/
http://element7digital.com.au/blog/9-proven-growth-hacks-to-increase-your-twitter-followers
https://www.twesocial.com/pricing.html
https://twiends.com/
- verify every single claim out there. This is the Internet. A lot of the time you do further investigation and the claims that a lot of people make no sense unless something 'strange has occurred'
- use your data and analytics... Some platforms don't offer internally so use alternate means... Fail fast so that you can succeed fast. Use A/B type testing. Else, all you're doing is burning away your company's money...
most useful social media platform business
follower analysis twitter
follower analysis instagram
- wanted to examine popular social media organisation accounts to see whether there were any commonalities. They post about half a dozen times a day. Often post uniquish content. Not necessarily always related 100% to the thing they're trying to hawk but there's definitely a bias there... Examples of this are NASA and National Geographic while Trump is much more heavily orientated towards their own hotel operations...
most popular twitter accounts
create follower bot twitter
social media accounts using bots
- This is the ultimate/basic bot that this relies on Selenium. It basically requires a seperate system (or VM running at all times). If had more time or resources I would have turned this on...
friend and like facebook bot python
facebot.py
facebook bot
- the other type of bot out there seems to be based on P2P pools that basically auto-like/follow one another. This is how I'm guessing some of the unrealistic bot systems out there work. It's basically bots following bots following other bots with people thrown in as often as possible...
auto follow pool twitter
- this is the only option/situation for low resource environments. I'm currently building something similar...
- don't get annoyed if your social media numbers 'oscillate' (namely, they seem to move up and down for no good reason in small numbers). That's basically 'bot on bot action'...
- note, from time to time significant changes in policies can have a massive impact on bots downstream...
- a lot of code out there just flat out doesn't work... Be prepared to build your own solutions if need be. It may actually be quicker then looking for a solution on the Internet
harvest email addresses from google
harvest email addresses from google github
send bulk email
- don't be afraid to use multiple bots to get the job done... Try not to use them at the same time though as it may get you banned...
auto follow back accounts twitter
twitter unfollow extension
- there are underground groups out there using their own codes, memes, etc... Figure it out and you can alienate or bring in large groups of people in one go...
- if you honestly want to know whether you're gaining genuine traffic the only way to do this is to keep things in house...
instagram bot
GrowBot Automator for Instagram
- don't get sucked into the 'Big Data' arena if you don't need to. As the intelligence services and large commercial organisations have discovered you need to sift/wade through huge amounts of data just to get at the data that you need. It's not efficient particularly for SME type firms...
- don't be too worried if your social media demographic seems strange (overly young, doesn't make sense, etc...). That's the nature of social media in general. Four figures followers will generally have you in a very tiny minority across most social media platforms... Note that if everyone is following best practice why do some pull out while most others fail? Think carefully about this... Don't be afraid to make mistakes as long as you learn them
we retweet news with hashtag
- for you to genuinely to be able to 'make a difference' in this area you need several months of work. This isn't the type of thing that you can simply knock over in a few days flat even if you do have a developer/marketing background. You need to study/analyse things for you to be able to make a genuine long term difference. Linear type growth is easy (standard bot configuration with standard material). Exponential/viral type growth requires decent content though... Learn not to be 'offended by the Internet' (you can't get angry or even with the Internet anyhow! LOL). Your product could be great but no one's interested because you haven't presented in a way that people find interesting relative to the competition, you may have done inadequate research into the market, you may have presented it poorly, you may have targeted the wrong people in your marketing campaign, etc... Just learn your way around. Hopefully, you'll figure out something eventually...

Random Stuff:
- as usual thanks to all of the individuals and groups who purchase and use my goods and services
- latest in science and technology
open kylix download
kylix 3 jessie debian
- latest in finance and politics
- latest in defense and intelligence
- latest in animal news
- latest in music and entertainment

Random Quotes:
- UNITED NATIONS –  North Korea says it can't pay nearly $184,000 in dues to the United Nations because of U.N. sanctions that prevent the transfer of funds from Pyongyang.
North Korea's U.N. Mission said sanctions imposed by the Security Council last August on the Foreign Trade Bank of the Democratic People's Republic of Korea, which is in charge of international transactions, have made payment impossible.
The mission said Ambassador Ja Song Nam met Undersecretary-General for Management Jan Beagle Friday to request the opening of "banking channels" so the DPRK can make its 2018 required payments for U.N. operations.
The United States banned the Foreign Trade Bank from the U.S. financial system in 2013 and the DPRK Mission blamed the Trump administration for spurring last August's U.N. sanctions against the bank.
- The experts I spoke to all stressed that Kim could devastate Seoul without even needing to use his weapons of mass destruction. The North Korean military has an enormous number of rocket launchers and artillery pieces within range of Seoul. The nonpartisan Congressional Research Service estimates that Kim could hammer the South Korean capital with an astonishing 10,000 rockets per minute — and that such a barrage could kill more than 300,000 South Koreans in the opening days of the conflict. That’s all without using a single nuclear, chemical, or biological weapon.
- Plenty of sensitive data has been found lying unsecured in S3 buckets, with the security firm UpGuard finding such stashes quite often.

UpGuard releases details of its finds on the Web regularly. It has found misconfigured S3 buckets leaking data from Paris-based brand marketing company Octoly, California data analytics firm Alteryx, credit repair service National Credit Federation, the NSA, the Pentagon, global corporate consulting and management firm Accenture, publisher Dow Jones, a Chicago voter database, a North Carolina security firm, and a contractor for the US National Republican Committe
- The British Home Office has announced that it has developed new technology that will be able to automatically detect what it calls "terrorist content" on all online platforms.

No date was given for rolling out the technology or what it would be called, but a statement, issued along with Home Secretary Amber Rudd (below), said tests had shown the technology in question could detect 94% of propaganda from the Islamic State group with 99.995% accuracy.

The statement also claimed that if the technology was used to analyse a million randomly selected videos, then only 50 would require additional review.

It said the technology had been developed jointly by the Home Office and artificial intelligence firm ASI Data Science and used "advanced machine learning to analyse the audio and visuals of a video to determine whether it could be Daesh propaganda". Daesh is another name for the Islamic State group.
- "While it may surprise you, it's not the least bit uncommon for jets to descend at what a pilot calls 'flight idle', with the engines run back to a zero-thrust condition," says pilot and author Patrick Smith in his book Cockpit Confidential. "They're still operating and powering crucial systems, but providing no push. You've been gliding many times without knowing it. It happens on just about every flight."

Smith explains that idle thrust is doing harder work than engines cut out altogether, but not by much. He likens it to coasting down a hill in a car with the clutch down.

Different aircraft have different glide ratios, meaning they will lose altitude at different rates, affecting how far they can fly without engine thrust. For example, if a plane has a lift to drag ratio of 10:1 then that means for every 10 miles (16 kilometres) of flight it loses one mile in altitude. Flying at a typical altitude of 36,000 feet (about 11 kilometres), an aircraft that loses both engines will be able to travel for another 70 miles (113 kilometres) before reaching the ground.
- "I am a person who would like to create opportunities for myself, because nobody could and would do that for me," Mr Meng said.
- So why do employers pay so much for people who’ve done them, or at least have got the certificates to show they once did them?

Caplan reckons it’s profiling, a bit like racial profiling, where police use the way someone looks as a rule of thumb to work out whether they are likely to commit a crime, or the profiling by insurance companies who use postcodes to tell them what to charge. It mightn’t be fair, but it's quick.

Seen that way, university is a sorting tool for employers, one they don’t pay for. It helps them identify characteristics that will be needed on the job but have nothing to do with what was learnt. One is intelligence. You need a certain amount to get enough marks to pass, whatever the subject. Another is conscientiousness. You need to apply yourself. And the third is conformity. Sane free-thinkers realise quickly there’s not a lot of point to what they are learning and drop out. Degrees certify IQ, the ability to knuckle down and a worker who won’t make trouble.

So they are great for employers and great for graduates, albeit at the cost of enormous wasted resources. Employers could get the same outcomes if the courses lasted for two years instead of four, or even one. Or if they administered tests themselves.

If Caplan’s right, we should be pushing politicians for less education rather than more, especially as the ageing of the population makes workers more scarce. My own company, Fairfax, is doing just that. It has taken on several truly excellent journalists precisely for the reason that they left university rather than see it through. They wanted to do the job rather than study it.

University isn’t for everyone, but life is. And it’s even better than university.
- The M27 was first being touted as an alternative to the M249 SAW. IMO this is a terrible idea and foolhardy. The M27 cannot provide the sustained fire the SAW can nor can it provide a high volume of fire. The standard load for the SAW gunner 600rounds. In contrast the standard load for a rifleman is 240 rounds (personally I always had 270 or 300 rounds). In a very real sense the M27 is the modern day equivalent to the BAR, which is I think a step in the wrong direction. If the US Military really wanted a suitable replacement to the M249 then I think they should take a look at the Negev and the new KAC LMG. 

I'm just an Army dog, but I think the USMC is onto something here. The paradigm is shifting. Volume of fire is great, but wasteful. Recent experience in Iraq and Asscrackistan has shown you can still lose the gunfight even if you're getting more rounds down range. With the 5.56mm NATO round (even the 70+ grain spec ops rounds) you reach a point of diminishing returns. You won't shoot through much and you won't have the range. Belt fed volume of fire is best served by 7.62mm NATO rounds found in the M240B. They will at least shoot through sandbags, walls, brick, mortar, heavy vegetation and more types of body armor. You get double the range too. The M27 allows them less volume, with much greater accuracy in a more common and lighter round. They like the M27 enough to use it as a DMR (not to be confused with sniper rifle). That has given birth to the M38, which is a DMR mod of a standard M27.
- Burke: The way you’re feeling right now … is why I have to believe in something bigger than me. Because if I didn’t, that powerlessness would eat me alive.
- Mr Vasta left the Caltex system last year after two years of financial hardship and emotional distress caused by his loss-making stores on Sydney’s suburban fringe.

Ahead of exiting the Caltex system, he had been on financial assistance from the oil refining and retail giant. At one point his debt relating to his stores had blown out to $355,000.

“I’m not a happy camper. We lost a lot of money,” he said.

As part of Caltex’s transition process, leases for franchises that end before 2020 will not be renewed. Those with leases ending after 2020 will receive some form of payout.

When asked if he thought it was fair to franchisees that Caltex would not be renewing leases after 2020, Mr Vasta said: “I think it’s wrong to ask if I think it’s fair. Fairness and Caltex do not go together."
- He told the interviewer that Indians were not encouraged to follow creative careers.

"The culture here is one of success based upon academic excellence, studying, learning, practising and having a good job and a great life. For upper India, not the lower. I see two Indias," he said.

"That's a lot like Singapore study, study, work hard and you get an MBA, you will have a Mercedes but where is the creativity? The creativity gets left out when your behaviour is too predictable and structured, everyone is similar.

"Look at a small country like New Zealand, the writers, singers, athletes, singers, athletes, it's a whole different world."

Asked about teaching students to code in schools, Wozniak said that while it was important, nobody should be taught to code before the age of 12.

"You don't get to a stage of symbolic reasoning until you are 12 years old. Some people get there early, but most people at 12, and that's why algebra can't be taught till you are that age. And programming can be taught only when you are ready for algebra," he said.

TPE SD2000i Inverter Petrol Power Generator Repair, Random Stuff, and More

- I recently got a TPE SD2000i Inverter Petrol Power Generator for a cheap price ~$50 at a local market. Apparently, it was purchased and us...