ramblings about spam, scalability, software development, the web

AWS Service SLAs

The goal of this page is to high-light the lack of coverage AWS provides for its services across different security factors. These limitations are not well-understood by many. Further, the “Y” fields are meant to indicate that this service has any capability for the relevant factor. In many cases, this is not full coverage for the service, or there are exceptions or special cases.

(tags: amazon aws services slas ops reliability)

Trek10 | From relational DB to single DynamoDB table: a step-by-step exploration

tl;dr:

Is modeling my relational database in a single DynamoDB table really a good idea? About a year ago, I wrote a fairly popular article called “Why DynamoDB isn’t for everyone”. Many of the technical criticisms of DynamoDB I put forth at that time (lack of operational controls such as backup/restore; a persistent problem with hot keys) have since been partially or fully resolved due to a truly awe-inspiring run of feature releases from the DynamoDB team. However, the central argument of that article remains valid: DynamoDB is a powerful tool when used properly, but if you don’t know what you’re doing it’s a deceptively user-friendly guide into madness. And the further you stray into esoteric applications like relational modeling, the more sure you’d better be that you know what you’re getting into. Especially with SQL-friendly “serverless” databases like Amazon Aurora hitting their stride, you have a lot of fully-managed options with a smaller learning curve.

(tags: dynamodb databases storage nosql sql relational aws relations)

Give Up the Ghost: A Backdoor by Another Name | Just Security

Now that GCHQ have asked for this, I suspect plenty of other government bodies around the world will be looking for similar.

They’re talking about adding a “feature” that would require the user’s device to selectively lie about whether it’s even employing end-to-end encryption, or whether it’s leaking the conversation content to a third (secret) party. Is the security code displayed by your device a mathematical representation of the two keys involved, or is it a straight-up lie? Furthermore, what’s to guarantee that the method used by governments to insert the “ghost” key into a conversation without alerting the users won’t be exploited by bad actors? Despite the GCHQ authors’ claim, the ghost will require vendors to disable the very features that give our communications systems their security guarantees in a way that fundamentally changes the trust relationship between a service provider and its users. Software and hardware companies will never be able to convincingly claim that they are being honest about what their applications and tools are doing, and users will have no good reason to believe them if they try. And, as we’ve seen already seen, GCHQ will not be the only agency in the world demanding such extraordinary access to billions of users’ software. Australia was quick to follow the UK’s lead, and we can expect to see similar demands, from Brazil and the European Union to Russia and China. (Note that this proposal would be unconstitutional were it proposed in the United States, which has strong protections against governments forcing actors to speak or lie on its behalf.) We must reject GCHQ’s newest “ghost” proposal for what it is: a mandated encryption backdoor that weakens the security properties of encrypted messaging systems and fundamentally compromises user trust.

(tags: crypto ghost gchq security backdoors uk)

FFmpeg, SOX, Pandoc and RSVG for AWS Lambda

OK-ish way to add dependencies to your Lambda containers:

The basic AWS Lambda container is quite constrained, and until recently it was relatively difficult to include additional binaries into Lambda functions. Lambda Layers make that easy. A Layer is a common piece of code that is attached to your Lambda runtime in the /opt directory. You can reuse it in many functions, and deploy it only once. Individual functions do not need to include the layer code in their deployment packages, which means that the resulting functions are smaller and deploy faster. For example, at MindMup, we use Pandoc to convert markdown files into Word documents. The actual lambda function code is only a few dozen lines of JavaScript, but before layers, each deployment of the function had to include the whole Pandoc binary, larger than 100 MB. With a layer, we can publish Pandoc only once, so we use significantly less overall space for Lambda function versions. Each code change now requires just a quick redeployment.

(tags: serverless lambda dependencies deployment packaging ops)

Allthefood

Decent new Dublin food blog — reviews and news. Like a Harbo-free version of Lovin’ Dublin

(tags: dublin food eating restaurants reviews)

We’ll Never Know Whether Monorepos Are Better

This makes sense to me where the monorepo-vs-polyrepo argument is concerned — it’s another religious war:

About six months after the project was declared “done” (but there was always more to do, more improvements to make to our homegrown dependency management solution), we had a retrospective meeting. The same engineers who had taken sides, for and against the project, were again assembled to discuss how it went. One of the main opponents went first. “Thank goodness we’re finally having this retrospective,” he said. “I think we can all see that this experiment has been a colossal failure and that it’s time for us to change course and roll back to monorepo.” “What do you mean?” one of the main multirepo advocates replied. “This was one of the best decisions we’ve ever made!” This really shocked me. We had access to all of the data you could possibly want to evaluate the decision. The same engineers working with the same codebase had seen what it was like in the monorepo model and the multirepo model. We knew exactly how much it had actually cost to switch. We had lived with the advantages and disadvantages of both models. But still we couldn’t come to an agreement. That retrospective taught me to be humble in my ambitions to “improve” engineering productivity. There’s no way to measure productivity in software, so there’s no way to know whether controversial, expensive “productivity enhancing” projects actually deliver on their promise, even in hindsight.

(tags: monorepo productivity dev engineering coding polyrepo)

I Accidentally Made Myself Lactose Intolerant With Whole30

A few years back, I had a nasty bout of food poisoning while travelling, which made me lactose-intolerant for several years. Sounds like this may be more common than you’d think, based on this article:

If you haven’t heard of Whole30, some information: It’s a month-long eating plan that aims to help followers hit “the reset button with your health, habits, and relationship with food.” For 30 days, you cut out soy, legumes, grains, sugars, alcohol, and, of course, dairy. [….] When you reach the end of the Whole30, you’re supposed to add the forbidden food groups back into your diet one at a time. The goal is to figure out which foods are making you feel sluggish, bloated, or just generally not great, so you can ostensibly keep on avoiding them forever. I didn’t do that part. I just jumped right back into eating what I wanted — but suddenly nothing was the same. That first bowl of ice cream I’d been looking forward to for weeks was quickly followed by sharp stomach pains and what can best be described (grossly, but accurately) as bubble gut. [….] The good news, according to gastrointestinal specialist Kim Barrett, a professor of medicine at the University of California, San Diego, is that I’m not crazy. The bad news is that dairy no longer agrees with my body’s biology. Turns out, it is possible to suddenly make yourself lactose intolerant. “To some extent, our ability to handle lactose is a use-it-or-lose-it phenomenon,” Barrett says. The body digests lactose — a disaccharide — by using lactase, an enzyme in the small intestine, to break it down into the monosaccharides glucose and galactose, which can then be absorbed. “If you don’t have the [lactose] substrate in the diet, you start to reduce the synthesis of the lactase enzyme to digest it,” Barrett explains. “After a period of completely excluding lactose from the diet, you may not have any of those digestive enzymes present.”

(tags: diet food lactose lactase intolerance whole30 milk cheese)

The Hydrogen Fuel Cell Scam

The need to have a massive and costly buildout of [hydrogen refuelling] infrastructure remains one of the most glaring problems of a supposed transition to a hydrogen economy. The cost of a single hydrogen fueling station is likely to be over $2 million. This is in contrast to the relatively modest $50,000 cost of deploying a high-speed battery-electric car charging station. Another factor here is the reality of putting into place an all-new infrastructure from scratch versus building off of an already existing electrical infrastructure that exists in every developed nation.  The fact that every building, garage, and lamp post in the US is already electrified means we simply have to add one final component to the existing and established network.

(tags: hydrogen green driving cars fuel fossil-fuels decarbonisation)

Don’t buy a 5G smartphone—at least, not for a while | Ars Technica

wow, 5G sounds like it’s going to be terrible

(tags: 5g 4g mobile-phones mobile tech hardware radio)

Risky business: linking _Toxoplasma gondii_ infection and entrepreneurship behaviours across individuals and countries | Proceedings of the Royal Society B: Biological Sciences

Using a saliva-based assay, we found that students (n = 1495) who tested IgG positive for Toxoplasma gondii exposure were 1.4× more likely to major in business and 1.7× more likely to have an emphasis in ‘management and entrepreneurship’ over other business-related emphases. Among professionals attending entrepreneurship events, T. gondii-positive individuals were 1.8× more likely to have started their own business compared with other attendees (n = 197). Finally, after synthesizing and combining country-level databases on T. gondii infection from the past 25 years with the Global Entrepreneurship Monitor of entrepreneurial activity, we found that infection prevalence was a consistent, positive predictor of entrepreneurial activity and intentions at the national scale, regardless of whether previously identified economic covariates were included. Nations with higher infection also had a lower fraction of respondents citing ‘fear of failure’ in inhibiting new business ventures. While correlational, these results highlight the linkage between parasitic infection and complex human behaviours, including those relevant to business, entrepreneurship and economic productivity.

(tags: science biology infection toxoplasmosis parasites humans behaviour entrepreneurs business brains economics)

“Thread on climate change & Irish agri lobbyists”

‘The massive levy income raised on beef & dairy gives the IFA a financial incentive to promote beef & dairy expansion and a disincentive to support any fundamental changes to Ireland’s agricultural model, regardless the public interest.’

(tags: dairy beef ireland farming climate-change climate lobbying government)

slides from “Distributed Log-Processing Design Workshop”, SRECon Americas 2018

Fantastic presentation discussing the kinds of design criteria used when architecting a large-scale data processing and storage service. Interesting to see some Google terminology, e.g. “dimensioning” — ballparking the expected scalability numbers, bandwidth, qps, and limits.

(tags: distributed-systems coding design architecture google photon logs log-storage slides srecon)

Zstandard: How Facebook increased compression speed – Facebook Code

very cool; particularly the high level of support for external-dictionary compression

(tags: facebook compression zstd zstandard dictionary-compression)

The Holiday Drink of 2018: Cognac Punch

Looks tasty, albeit pricey and strong:

London Punch House Punch By David Wondrich Ingredients: 4 Lemons 6 oz (3/4 cup) White sugar 20 oz Hennessy Privilege VSOP Cognac 6 oz Jamaican rum 1 quart (4 cups) Cold water Glass: Punch cup Garnish: Freshly grated nutmeg Directions: The day before you serve the punch, fill a quart container with water and put it in the freezer. Peel 4 lemons in long spirals and put the peels into a mason jar with 6 oz (3/4 cup) of white sugar. Seal the jar, shake it and let sit overnight. Two hours before you serve the punch, unseal the mason jar, add 6 oz (3/4) cup fresh-squeezed, strained lemon juice, reseal and shake until all the sugar has dissolved. Refrigerate. To assemble the punch, unmold the ice block and place it in a 1-gallon punch bowl. Shake the contents of the mason jar and pour it into the punch bowl unstrained, peels and all. Add 20 oz (2-and-a-half cups) Hennessy Privilege VSOP Cognac and 6 oz (3/4 cup) Jamaican rum. Stir. Add 1 quart (4 cups) cold water. Stir again and grate nutmeg on top. Using a punch ladle, drape a few ends of the spiral lemon peels over the rim of the bowl. Grate nutmeg over the top.

(tags: punch cognac hennessy recipes rum)

CRDTs in Production

at PayPal (via Carlos Baquero)

(tags: via:xmal crdts distributed-systems distcomp prod coding data-structures)

Remove Background from Image

A fully automated background-removal tool. Nicely done

(tags: images editing tools background graphics)

Google used a Baidu front-end to scrape user searches without consent

The engineers used the data they pulled from [acquired Baidu front-end site] 265.com to learn about the kinds of things that people located in mainland China routinely search for in Mandarin. This helped them to build a prototype of Dragonfly. The engineers used the sample queries from 265.com, for instance, to review lists of websites Chinese people would see if they typed the same word or phrase into Google. They then used a tool they called “BeaconTower” to check whether any websites in the Google search results would be blocked by China’s internet censorship system, known as the Great Firewall. Through this process, the engineers compiled a list of thousands of banned websites, which they integrated into the Dragonfly search platform so that it would purge links to websites prohibited in China, such as those of the online encyclopedia Wikipedia and British news broadcaster BBC. Under normal company protocol, analysis of people’s search queries is subject to tight constraints and should be reviewed by the company’s privacy staff, whose job is to safeguard user rights. But the privacy team only found out about the 265.com data access after The Intercept revealed it, and were “really pissed,” according to one Google source.

(tags: china search tech google privacy baidu interception censorship great-firewall dragonfly)

awslabs/amazon-kinesis-scaling-utils

The Kinesis Scaling Utility is designed to give you the ability to scale Amazon Kinesis Streams in the same way that you scale EC2 Auto Scaling groups – up or down by a count or as a percentage of the total fleet. You can also simply scale to an exact number of Shards. There is no requirement for you to manage the allocation of the keyspace to Shards when using this API, as it is done automatically. You can also deploy the Web Archive to a Java Application Server, and allow Scaling Utils to automatically manage the number of Shards in the Stream based on the observed PUT or GET rate of the stream.

(tags: kinesis scaling scalability shards sharding ops)

The “Bart” — sudden hundreds-of-Bitcoin pumps or dumps, to burn the margin traders

Finance journalists need to stop treating crypto as an efficient market that responds to concerns. It’s a thinly-traded unregulated playground for whales, out to wreck the margin traders. A $400 dip in fifteen minutes is not a “market signal” — it’s a deliberate dump to manipulate the price. Though there’s still downward pressure on the price — all the suckers from the bubble have gone home, so they’re not buying … but the miners still have to sell coins for actual money to pay for their electricity. And even more so now that the price of mining one bitcoin is at — or above — what you could get for selling that bitcoin. So one minute you’ll see a sudden $100 increase in the price that cost 130 BTC of dollars — and those are actual dollars going in — followed the next minute by a matching $100 drop that came from selling only 30 BTC. It’s much easier to drop the price than raise it.

(tags: investing bitcoin via:harikunzru cryptocurrency pump-and-dump barts finance margin-traders currencies fraud)

GDPR Subject Access Request helper for Dublin Bus vs bike incidents

‘A new tool has been created to assist people who want to request video footage of incidents involving Dublin Bus under GDPR rules.’

(tags: dublin-bus bus bike cycling dublin cameras gdpr sars)

8 Myths of Abortion in Ireland

from the Abortion Rights Campaign’s “8 Days – 8 Reasons to Repeal the 8th” myth-busting series in 2014.

(tags: arc abortion ireland politics mythbusters repealthe8th)

Notebookcheck’s Top 10 Tablets under 250 Euros

a decent list of low-end tablets, given the season

(tags: tablets devices hardware android gadgets xmas)

A primer on privacy as “contextual integrity”

A primer on privacy as “contextual integrity” and why privacy notices on mobile platforms (both Android and iOS) are insufficient for attaining informed consent. If your doctor asked for permission to collect your medical history, you would probably say yes. However, if that doctor asked to collect your medical history to give to marketers for advertising purposes, you would probably decline. The difference is, in the first case, you’re making assumptions about how the data will be used based on who is making the request. Knowing just the type of data requested and the requester are insufficient to make an informed decision: people also consider the purpose and other constraints (e.g., will data be resold, stored securely, etc.), which are equally important factors.

(tags: privacy context understanding mobile data-protection permission opt-in)

Makisu

Uber’s Fast, Reliable Docker Image Builder for Apache Mesos and Kubernetes.

we built our own image building tool, Makisu, a solution that allows for more flexible, faster container image building at scale. Specifically, Makisu: requires no elevated privileges, making the build process portable. uses a distributed layer cache to improve performance across a build cluster. provides flexible layer generation, preventing unnecessary files in images. is Docker-compatible, supporting multi-stage builds and common build commands.

(tags: makisu docker containers ops build mesos kubernetes building)

Billionaires Are the Leading Cause of Climate Change

That’s largely because there is no “free market” incentive to prevent disaster. An economic environment where a company is only considered viable if it’s constantly expanding and increasing its production can’t be expected to pump its own brakes over something as trivial as pending global catastrophe. Instead, market logic dictates that rather than take the financial hit that comes with cutting profits, it’s more reasonable to find a way to make money off the boiling ocean. Nothing illustrates this phenomenon better than the burgeoning climate-change investment industry. According to Bloomberg, investors are looking to make money off of everything from revamped food production to hotels for people fleeing increasingly hurricane-ravaged areas. A top JP Morgan Asset investment strategist advised clients that sea-level rise was so inevitable that there was likely a lot of opportunity for investing in sea-wall construction.

(tags: capitalism environment politics future climate-change doom)

Irish agricultural CO2 emissions actually INCREASED by 2.9% last year

Irish EPA: agriculture greenhouse gas emissions increased by 2.9% in 2017 “The most significant drivers are higher dairy cow numbers (+3.1%) which reflects national plans to expand milk production”

(tags: epa ireland co2 greenhouse-gases emissions green farming agriculture)

“Bring home the bacon” debate is a case study in how stupid we’ve all become

Well said, that Carl Kinsella.

How are we supposed to focus on the things that matter when media organisations refuse to? Ivan Yates and Matt Cooper ran a segment on whether or not we’ve reached political correctness. Niall Boylan had a discussion about whether or not it’s okay to ban phrases that nobody really wants to ban [as PETA has suggested]. As a matter of pure coincidence, other reports from outlets like the Daily Mail and Today FM reported that the HSE was going to ban its employees from using phrases like “love,” “pet” or “dear.” That wasn’t true either. Again, it was one line in a huge report that asked the question as to whether or not it was the best course of action to use such terms. It wasn’t mentioned again. Unlike phrases about meat, this actually matters. This is the kind of thing that could actually change how people are treated in hospital. But again, it wasn’t true. The public – which has lost its capacity to verify what is true and what is not — was being lied to by the media. That’s how it works now.

(tags: media clickbait fake-news news today-fm daily-mail hse niall-boylan peta controversy political-correctness ivan-yates matt-cooper)

Rudy Giuliani doesn’t understand how links work

As waxy noted: ‘this might be funny if he wasn’t Trump’s cybersecurity advisor’.

Twitter allowed someone to invade my text with a disgusting anti-President message. The same thing-period no space-occurred later and it didn’t happen. Don’t tell me they are not committed cardcarrying anti-Trumpers. Time Magazine also may fit that description. FAIRNESS PLEASE

(tags: edge-cases bugs twitter regexps regular-expressions links urls us-politics trump rudy-giuliani security funny)

3D models by DH_Age Sheela-na-Gig3D Project (@DH_Age) – Sketchfab

These are fantastic — 3D scans of Sheela-na-Gig carvings around Ireland from 3D Sheela, an Irish based research initiative ‘focusing on the digital documentation and analysis of Ireland’s Sheela-na-Gig catalogue’ (NSFW)

(tags: 3d sheela-na-gigs history carving nsfw models photogrammetry)

Chester Beatty Digital Collections

‘Explore online access to our remarkable treasures, through this searchable database of digitised artworks and manuscripts’, from the Chester Beatty museum’s collection. Licensing isn’t fully open though — ‘Images and PDF’s are provided for personal research and scholarship.’

(tags: chester-beatty museums history archaeology artifacts art manuscripts)

Convert an RSS feed to ActivityPub

for autoposting to Mastodon or similar. Via Nelson

(tags: via:nelson mastodon rss blogging scripts)

A Serverless Journey: AWS Lambda Under the Hood (SRV409-R1) – AWS re:…

Marc Brooker on Lambda internals

(tags: lambda aws marc-brooker ops slides)

Nice heatmap visualisation of mortality ratios between men and women

‘Age and period specific mortality ratios of females and males (a “Lexis-surface” in demographers parlance)’ (via apenwarr)

(tags: via:apenwarr mortality heatmaps dataviz visualisation colouring lexis-surfaces)

House Democrats Sent A New Letter To Jeff Bezos About Amazon’s Facial Recognition Tool

The House Democrats’ questions focus largely on possible embedded bias in Amazon Rekognition, including how the tool’s accuracy breaks down by race, gender, ethnicity, and age. Also of particular concern is whether Amazon will build privacy protections into its facial recognition system and how it will ensure it is not abused for secret government surveillance. [….] Meanwhile, Jeff Bezos has yet to address mounting criticism of Amazon’s Rekognition technology by Amazon employees, shareholders, and civil rights groups. In November, Amazon executives defended the company’s controversial facial recognition technology at an all-hands staff meeting after employees raised civil rights concerns about the tech’s potential misuse. “It’s hard to trust that harm and abuse can be prevented if it is only post-mortem and through the Terms of Service,” an Amazon employee who requested anonymity told BuzzFeed News at the time.

(tags: rekognition aws privacy data-protection surveillance amazon us-politics civil-rights)

Google Tried to Patent My Work After a Job Interview

I looked up the patent application and luckily, this time the patent application was still being reviewed by the patent examiner.  It had not issued! The provisional was filed August 29, 2014, months after my first interview and visit back in March 2014.  Two of the inventors listed were the same people who had interviewed me. 

(tags: patents software-patents google dirty-tricks interviewing ip mit medialab paper jie-qi)

Crowdfunding Backer Patented My Project

I had thought thought all along that if we published everything openly, it wouldn’t be possible for someone else to patent stuff that’s already all over the web.  But I was wrong. Despite tons of prior art out on the web, in academic research papers and even for sale that are LED stickers, the patent examiner missed it and deemed the LED sticker patent “new” and “non-obvious.”   How could that happen?  The sad truth is that patents are approved all that time that probably shouldn’t be. […] Can’t you invalidate the patent? Unfortunately, once a patent gets issued things get much more complicated and expensive. The cheapest option would be for us to go to the USPTO with our prior art list and invalidate the patent though a process called an inter partes review (IPR).  But such a process typically costs between $300,000 and $600,000 to file due to legal fees. In fact, it’s much more than the cost of getting the patent in the first place! So for now, we’ve decided not to go down this route.

(tags: patents led-stickers uspto prior-art invention crowdfunding)

PRDD – Performance-Review Driven Development

‘If the way to get promoted is to launch a shiny new product, then your most senior people will be the best at finding shiny new products to launch, even if that’s not the right technical decision to make.’ (from a newsy thread about Twitter’s latest messaging system switch)

(tags: newsy messaging infrastructure twitter kafka pubsub ops architecture prdd performance-reviews)

Presto Performance for Ad Hoc Workloads on AWS Instance Types

good benchmark/review of instance types from Qubole

(tags: qubole presto performance benchmarks ops aws instances ec2)

Party Parrot as a Service

Enter an image URL and it’ll generate an animated GIF of the party parrot version

(tags: party-parrot gifs funny slack emojis)

event-stream vulnerability explained – Zach Schneider

This was an incredibly clever attack, very reminiscent of this blog post from January about how a similar attack might work. The attacker covered their tracks well — the code and commit log on GitHub all tell an innocuous and fairly common story (a new maintainer joins a project, adds a feature, and then tweaks the implementation of their feature a bit). Other than the warning signs about flatmap-stream (new package, no contributors or download activity), the attack was virtually undetectable. And indeed, it wasn’t discovered for over two months — it was only found because the attacker made a tiny mistake and used the deprecated crypto.createDecipher rather than crypto.createDecipheriv, which raised a suspicious deprecation warning in another library that consumes event-stream. Unfortunately, this genre of attack isn’t going away anytime soon. JavaScript is the most popular language right now and it’s not really close, meaning it will continue to be an attractive target for hackers. JavaScript also has relatively few standard-library convenience features compared to other languages, which encourages developers to import them from npm packages instead — this, along with other cultural factors, means that JavaScript projects tend to have massive dependency trees.

(tags: npm malware bitcoin security javascript event-stream flatmap-stream hacks)

Leeroy Jenkins was taking the piss

I had no idea!

“We didn’t think anyone would believe it was real, we thought it was so obviously satire.”

(tags: leeroy-jenkins warcraft gaming memes satire videos at-least-i-have-chicken)

cryptoshredding

A way to handle immutable blockchains and GDPR: “have an encryption key for each user (stored outside of this ledger) and encrypt all PII with that key. Throw away the key if the user wants you to delete their data.”

(tags: pii gdpr privacy data-protection crypto cryptoshredding deletion coding)

flatmap-stream NPM package backdoor incident

Good twitter thread with background on the incident. 2,000,000 downloads per week, used by many other core libs. It appears the attacker persuaded the (overloaded) legit maintainer to hand over ownership then backdoored the package in order to attack copay-dash, a cryptocurrency wallet app.

(tags: cryptocurrency npm packages open-source twitter flatmap-stream packaging security backdoors)

Antoin O Lachtnain on Twitter regarding GDPR and DNA kits

‘It provides legal protections. But the problem is that those protections may not be practically effective. The other problem is the effect that you, as a consenting adult may have on other people who are related to you but who haven’t consented.’ — useful thread

(tags: gdpr privacy data-protection dna)

boundary-layer

Declarative Airflow Workflows in YAML, from Etsy

(tags: airflow python batch cron etl)

Linkerd 2.0

The 2.0 release of Linkerd brings two very significant changes. First, we’ve completely rewritten Linkerd to be orders of magnitude faster and smaller than Linkerd 1.x. Linkerd 2.0’s data plane is comprised of ultralight Rust proxies which consume around 10mb of RSS and have a p99 latency of <1ms. Linkerd’s minimalist control plane (written in Go) is similarly designed for speed and low resource footprint. Second, with the 2.0 release, Linkerd moves beyond the service mesh model to be something not more, but less: Linkerd 2.0 at its core is a service sidecar, running on a single service without requiring cluster-wide installation. This means that if you’re a developer or service owner who doesn’t have access to the whole Kubernetes cluster, you can run Linkerd on your service and get: Instant Grafana dashboards of your service’s success rates, latencies, and throughput; A topology graph of incoming and outgoing dependencies; A live view of requests being made to your service; Improved, latency-aware load balancing; … and much more.

(tags: linkerd mesh networking services architecture sidecars)

MuMufication

the act of having a small portion of your cremated remains fired in a Brickof Mu. MuMufied is what you will be after the act of MuMufication has been carried out. What you get in the here and now is a Brick of Mu and a signed and stamped Certificate of MuMufication. What you get after you die is 23 grams of yours cremated remains fired in your Brick of Mu, which will then be laid to rest on The People’s Pyramid come the following Toxteth Day of the Dead on 23rd of November.

(tags: klf mu mumufication 23 eris discordianism religion death toxteth liverpool cremation pyramids)

Introducing Predictive Scaling for Amazon EC2 in AWS Auto Scaling

Predictive Scaling predicts future traffic based on daily and weekly trends, including regularly-occurring spikes, and provisions the right number of EC2 instances in advance of anticipated changes. Provisioning the capacity just in time for an impending load change makes Auto Scaling faster than ever before. Predictive Scaling’s machine learning algorithms detect changes in daily and weekly patterns, automatically adjusting their forecasts. This removes the need for manual adjustment of Auto Scaling parameters over time, making Auto Scaling simpler to configure and consume. Auto Scaling enhanced with Predictive Scaling delivers faster, simpler, and more accurate capacity provisioning to our customers.

(tags: aws amazon scaling autoscaling predictive-scaling ml ec2 asg)

Awful AI

Artificial intelligence in its current state is unfair, easily susceptible to attacks and notoriously difficult to control. Nevertheless, more and more concerning uses of AI technology are appearing in the wild. This list aims to track all of them. We hope that Awful AI can be a platform to spur discussion for the development of possible contestational technology (to fight back!).

(tags: ai algorithms ethics technology machine-learning)

Some notes about HTTP/3

Robert Graham from ErrataSec on QUIC aka HTTP/3: ‘Google (pbuh) has both the most popular web browser (Chrome) and the two most popular websites (#1 Google.com #2 Youtube.com). Therefore, they are in control of future web protocol development.’ Faster connection setup and latency; better bandwidth negotiation when using multiplexing; user-mode stacks by building on UDP and using recvmmsg(); and better mobile support for roaming IPs.

(tags: google http3 quic protocols ip)

Traditional Chinese medicine origins: Mao invented it but didn’t believe in it

Mikulski and the rest of the Senate may be surprised to learn that they were repeating 60-year-old justifications of Chinese medicine put forward by Chairman Mao. Unlike Mikulski, however, Mao was under no illusion that Chinese medicine—a key component of naturopathic education—actually worked. In The Private Life of Chairman Mao, Li Zhisui, one of Mao’s personal physicians, recounts a conversation they had on the subject. Trained as an M.D. in Western medicine, Li admitted to being baffled by ancient Chinese medical books, especially their theories relating to the five elements. It turns out his employer also found them implausible. 

(tags: medicine tcm mao history china health naturopathy)

Java’s ByteBuffer native memory “leak”

Well this is suboptimal:

The Java NIO APIs use ByteBuffers as the source and destination of I/O calls, and come in two flavours. Heap ByteBuffers wrap a byte[] array, allocated in the garbage collected Java heap. Direct ByteBuffers wrap memory allocated outside the Java heap using malloc. Only “native” memory can be passed to operating system calls, so it won’t be moved by the garbage collector. This means that when you use a heap ByteBuffer for I/O, it is copied into a temporary direct ByteBuffer. The JDK caches one temporary buffer per thread, without any memory limits. As a result, if you call I/O methods with large heap ByteBuffers from multiple threads, your process can use a huge amount of additional native memory, which looks like a native memory leak. This can cause your process to unexpectedly run into memory limits and get killed.

(tags: jvm performance java memory leaks bytebuffers netty threads coding bugs)

The Time Our Provider Screwed Us

Good talk (with transcript) from Paul Biggar about what happened when CircleCI had a massive security incident, and how Jesse Robbins helped them do incident response correctly. ‘On the left, Jesse pointed out that we needed an incident commander. That’s me, Paul. And this is very good, because I was a big proponent, I think lots of were around the 2013 mark, of flat organizational structures, and so I hadn’t really got a handle of this whole being in charge thing. The fact that someone else came in and said, “No, no, no, you are in charge”: extremely useful. And he also laid out the order of our priorities. Number one priority; safety of customers. Number two priority: communicate with customers. Number three priority: recovery of service. I think a reasonable person could have put those in a different order, especially under the pressure and time constraints of the potential company-ending situation. So I was very happy to have those in order. If this is ever going to happen to you, I’d memorize them, maybe put it on an index card in your pocket, in case this ever happens. The last thing he said is to make sure that we log everything, that we go slow, and that we code review and communicate. His point there is that if we’re going to bring our site back up, if we’re going to do all the things that we need to do in order to save our business and do the right thing for our customers and all that, we can’t be making quick, bad decisions. You can’t just upload whatever code is on your computer now, because I have to do this now, I have to fix it. So we set up a Slack channel … This was pre-Slack; it was a HipChat channel, where all of our communications went. Every single communication that we had about this went in that chatroom. Which came in extremely useful the next day, when I had to write a blog post that detailed exactly what had happened and all the steps that we did to fix it and remediate this, and I had an exact time stamps of all the things that had happened.’

(tags: incidents incident-response paul-biggar circleci security communication outages)

Deep learning can “discover” new knowledge from scans/images

Amazing paper:

Here, we show that deep learning can extract new knowledge from retinal fundus images. Using deep-learning models trained on data from 284,335 patients and validated on two independent datasets of 12,026 and 999 patients, we predicted cardiovascular risk factors not previously thought to be present or quantifiable in retinal images, such as age (mean absolute error within 3.26 years), gender (area under the receiver operating characteristic curve (AUC)?=?0.97), smoking status (AUC?=?0.71), systolic blood pressure (mean absolute error within 11.23?mmHg) and major adverse cardiac events (AUC?=?0.70). We also show that the trained deep-learning models used anatomical features, such as the optic disc or blood vessels, to generate each prediction.

(tags: deep-learning data analysis ml machine-learning health medicine papers)

OpsMop

‘a next-generation, no-compromise automation system’.

Uses: Web-scale configuration management of all Linux/Unix systems; Application deployment; Immutable systems build definition; Maintaining stateful services such as database and messaging platforms; Automating one-off tasks & processes; Deployment and management of the undercloud. Features: Python 3 DSL; Declarative resource model with imperative capabilities; Type / Provider plugin seperation; Implicit ordering (with handler notification); Formalized “Plan” vs “Apply” evaluation stages; Early validation prior to runtime; Programatically scoped variables; Strong object-orientation

(tags: opsmop ops configuration-management deployment build)

The JVM in Docker 2018

Later JDK versions have made it far easier to run a JVM application in a Linux container. The memory support means that if you relied on JVM ergonomics before than you can do the same inside a container where as previously you had to override all memory related settings. The CPU support for containers needs to be carefully evaluated for your application and environment. If you’ve previously set low cpu_shares in environments like Kubernetes to increase utilisation while relying on using up unused cycles then you might get a shock.

(tags: jvm docker kubernetes linux containers ops)

This Week’s Giza Pyramid Discovery Was Already Built Into Assassin’s Creed Origins

As a kid, I was Egypt-obsessed — this game is hitting the spot

(tags: assassins-creed egypt history pyramids architecture)

Tuning Spark Back Pressure by Simulation

Interesting, Spark uses a PID controller algorithm to manage backpressure:

Spark back pressure, which can be enabled by setting spark.streaming.backpressure.enabled=true, will dynamically resize batches so as to avoid queue build up. It is implemented using a Proportional Integral Derivative (PID) algorithm. This algorithm has some interesting properties, including the lack of guarantee of a stable fixed point. This can manifest itself not just in transient overshoot, but in a batch size oscillating around a (potentially optimal) constant throughput. The overshoot incurs latency; the undershoot costs throughput. Catastrophic overshoot leading to OOM is possible in degenerate circumstances (you need to choose the parameters quite deviously to cause this to happen). Having witnessed undershoot and slow recovery in production streaming jobs, I decided to investigate further by testing the algorithm with a simulator.

(tags: backpressure streaming queueing pid-controllers algorithms congestion-control)

New – EC2 Auto Scaling Groups With Multiple Instance Types & Purchase Options | AWS News Blog

Basically getting EC2 Fleet’s featureset into ASGs, good news

(tags: ec2 fleet asg ops architecture cost-control)

SpamAssassin is back [LWN.net]

The SpamAssassin 3.4.2 release was the first from that project in well over three years. At the 2018 Open Source Summit Europe, Giovanni Bechis talked about that release and those that will be coming in the near future. It would seem that, after an extended period of quiet, the SpamAssassin project is back and has rededicated itself to the task of keeping junk out of our inboxes.

(tags: spamassassin open-source oss anti-spam)

Google ‘betrays patient trust’ with DeepMind Health move | Technology | The Guardian

Now that Streams is a Google product itself, that promise appears to have been broken, says privacy researcher Julia Powles: “Making this about semantics is a sleight of hand. DeepMind said it would never connect Streams with Google. The whole Streams app is now a Google product. That is an atrocious breach of trust, for an already beleaguered product.” A DeepMind spokesperson emphasised that the core of the promise remains intact: “All patient data remains under our partners’ strict control, and all decisions about its use lie with them. This data remains subject to strict audit and access controls and its processing remains subject to both our contracts and data protection legislation. The move to Google does not affect this.”

(tags: google deepmind health nhs data-protection privacy healthcare)

Jeff Bezos is wrong, tech workers are not bullies

I decided to leave my job as a staff engineer at Google because of Project Maven, and because I believe that the artificial intelligence ethical guidelines they published afterwards were not strict enough: they allowed surveillance within “internationally accepted norms”. I am now joining forces with current and former Google employees who also opposed Maven and the Dragonfly search engine. We do not wish to be complicit in human rights violations and we believe that workers, and the public, deserve a voice. We support employees at Amazon, Microsoft, Salesforce, McKinsey and Deloitte who have similarly stood up to their employers. We also have a right to not contribute to killing. Most workers at Google or Amazon did not join those companies to work on military applications. Both companies are international employers with engineering offices across the world, and many of their workers are neither US citizens nor residents. I worked as an engineer in Google’s European headquarters in Dublin, Ireland. To me, the US military is not our military (as Google Cloud chief executive Diane Greene referred to it in a blog post), nor is it a force we should automatically support as a matter of patriotism. As an engineer, I believe it is my responsibility to speak up for human rights and accountable decision making. As an industry, we in technology cannot compromise our principles or allow ourselves to be bullied by billionaires who stand to be enriched by our silence.

(tags: project-maven dragonfly google amazon surveillance us-politics politics ai silicon-valley ethics work life)

HTTP-over-QUIC to be renamed HTTP/3

Decent newsy comment thread about HTTP/3, QUIC, and how the modern internet treats IP protocols

(tags: ip protocols http http3 quic networking internet newsy)

Brits getting into the online-disinfo game

Disinformation and deception have been a part of warfare for thousands of years, but across the world, something new was starting to happen. Information has long been used to support combat operations, but now combat was seen to taking place primarily, sometimes exclusively, through it. From being a tool of warfare, each military began to realise that the struggle with, over and through information was what war itself actually was about. And it wasn’t confined to Russia, China or anyone else. A global informational struggle has broken out. Dozens of countries are already doing it. And these are just the campaigns that we know about.

(tags: disinfo fake-news uk british-army military memes infowar)

Some notes on running new software in production

This is really good — how to approach new infrastructure/software dependencies in production with reliability and uptime in mind. (via Tony Finch)

(tags: reliability uptime slas kubernetes envoy outages runbooks ops)

Specification gaming examples in AI

A good list of examples where machine learning systems “figure out” how to cheat their fitness function, e.g.:

‘Creatures bred for speed grow really tall and generate high velocities by falling over’

(tags: ai funny humor spreadsheets machine-learning ml fitness-functions)

What if the Placebo Effect Isn’t a Trick? – The New York Times

It is not possible to assay levels of COMT directly in a living brain, but there is a snippet of the genome called rs4680 that governs the production of the enzyme, and that varies from one person to another: One variant predicts low levels of COMT, while another predicts high levels. When Hall analyzed the I.B.S. patients’ DNA, she found a distinct trend. Those with the high-COMT variant had the weakest placebo responses, and those with the opposite variant had the strongest. These effects were compounded by the amount of interaction each patient got: For instance, low-COMT, high-interaction patients fared best of all, but the low-COMT subjects who were placed in the no-treatment group did worse than the other genotypes in that group. They were, in other words, more sensitive to the impact of the relationship with the healer. The discovery of this genetic correlation to placebo response set Hall off on a continuing effort to identify the biochemical ensemble she calls the placebome — the term reflecting her belief that it will one day take its place among the other important “-omes” of medical science, from the genome to the microbiome. The rs4680 gene snippet is one of a group that governs the production of COMT, and COMT is one of a number of enzymes that determine levels of catecholamines, a group of brain chemicals that includes dopamine and epinephrine. (Low COMT tends to mean higher levels of dopamine, and vice versa.) Hall points out that the catecholamines are associated with stress, as well as with reward and good feeling, which bolsters the possibility that the placebome plays an important role in illness and health, especially in the chronic, stress-related conditions that are most susceptible to placebo effects.

(tags: placebo comt health healthcare medicine enzymes brain)

Deadlines, lies and videotape: The tale of a gRPC bug

HostedGraphite decided to use gRPC as an internal inter-service protocol and ran into a basic protocol bug — it does not default to using an application-level keepalive on the TCP channel so can block indefinitely if sending-side buffers fill up. Always use application-level keepalives and don’t trust TCP

(tags: tcp protocols keepalive grpc rpc architecture networking)

Sci-Fi Writer Greg Egan and 4chan anon Math Whiz Advance Permutation Problem | Quanta Magazine

On September 16, 2011, an anime fan posted a math question to the online bulletin board 4chan about the cult classic television series ‘The Melancholy of Haruhi Suzumiya’. Season one of the show, which involves time travel, had originally aired in non-chronological order, and a re-broadcast and a DVD version had each further rearranged the episodes. Fans were arguing online about the best order to watch the episodes, and the 4chan poster wondered: If viewers wanted to see the series in every possible order, what is the shortest list of episodes they’d have to watch? In less than an hour, an anonymous person offered an answer — not a complete solution, but a lower bound on the number of episodes required. The argument, which covered series with any number of episodes, showed that for the 14-episode first season of Haruhi, viewers would have to watch at least 93,884,313,611 episodes to see all possible orderings. “Please look over [the proof] for any loopholes I might have missed,” the anonymous poster wrote. The proof slipped under the radar of the mathematics community for seven years — apparently only one professional mathematician spotted it at the time, and he didn’t check it carefully. But in a plot twist last month, the Australian science fiction novelist Greg Egan proved a new upper bound on the number of episodes required. Egan’s discovery renewed interest in the problem and drew attention to the lower bound posted anonymously in 2011. Both proofs are now being hailed as significant advances on a puzzle mathematicians have been studying for at least 25 years.

(tags: mathematics internet math greg-egan anime bizarre 4chan superpermutation permutation proofs)

How do you populate your development databases?

Lots of comments pro/anti copying from production

(tags: database data testing system-tests dev)

Welcome To The Dystopia: People Are Arguing Whether This Trump Press Conference Video Is Doctored Or Not

To sum it up: A historically unreliable narrator who works for a conspiracy website tweets out a video in order to show alleged bad behavior on the part of a journalist. The clip goes viral. The White House picks up and disseminates that video and uses it as proof to ban the journalist from reporting at the White House. Outraged journalists decry the White House’s use of a video taken from a historically unreliable narrator. Then, users attempt to debunk the video as “actual fake news.” Others, unclear if the video is fake, urge caution, suggesting the media may be jumping the gun. An argument breaks out over the intricate technical details of doctoring a clip. The entire ordeal is a near perfect example of a scenario disinformation experts have predicted and warned of, where the very threat of video manipulation can lead to a blurring of reality. “These technological underpinnings [of AI and photoshop, and editing programs lead] to the increasing erosion of trust,” computational propaganda researcher Renee DiResta told BuzzFeed News in early 2018. “It makes it possible to cast aspersions on whether videos — or advocacy for that matter — are real.”

(tags: twitter disinfo disinformation buzzfeed video reality fake-news photoshop)

wifiscan.py

A neat kragen hack:

Report wireless signal strength using audio synthesis. This quick kludge is useful for figuring out where the dead spots in your wireless network coverage are. They’re where the pitch goes high and maybe you start getting clicks from audio buffer underruns. Example audio output can be found at . This is based on a brilliant hack by Seth Schoen, in which he generated a tone controlled by the signal strength indication from his wireless card in order to get a sort of 21st-century software Theremin. This adds some envelope modulation so that it’s useful as a network diagnostic tool rather than a musical instrument. It depends on PulseAudio and the interface to the usual set of Linux commands: pacat, ping, and iwconfig. (I should really just read /proc/net/wireless instead of depending on iwconfig.)

(tags: kragen wifi scanning audio sound signal linux)

Underground Utility Colour Codes Explained

Underground utility colour codes are used to differentiate and identify underground utilities to protect it from damage during excavation. There are different types of utilities and in order to tell them apart coloured lines, flags or sometimes both are used. They help mark the location and indicate the type of utility that is buried underground.

(tags: utilities power underground excavation digging repair urban)

Meet the Irish master codebreaker you never heard of

In 1943, one of Nazi Germany’s most notorious communication codes was broken by a mild-mannered librarian and family man from Abbeyfeale in west Limerick, Richard Hayes. His day job was director of the National Library of Ireland, but during the Emergency, he secretly led a covert team of cryptanalysts, working feverishly on the infamous “Görtz Cipher” – a fiendish Nazi code that had stumped many of the greatest code-breaking minds at Bletchley Park, the centre of British wartime cryptography. Astonishingly, the feat was accomplished not in one of the huts at Bletchley Park, but in a now derelict building colloquially known as “The Red House” on Montpellier Hill in Dublin. If this wasn’t an amazing enough feat, Hayes also broke a complex microdot enciphering system which had baffled the American OSS, as well an enciphering system used by the dreaded Sicherheitsdienst (or SD), the intelligence section of the SS. The breaking of the latter code was instrumental in the Allied victory at the Battle of the Bulge. Few individuals can be said to have altered the course of the second World War on their own, but Hayes was one such individual – and he did it not with a gun, but with a detailed knowledge of the complex algorithms and mathematical permutations involved in the art of cryptography.

(tags: via:colmmacc crypto codes hermann-gortz wwii dublin ireland richard-hayes cryptanalysis)

Physical spam using cheap Bluetooth beacons spotted in the wild

Buzz Andersen on Twitter: “Crazy thing: my friends who run a high traffic coffee shop just found this thing under their bar. It turns out to be a super cheap Chinese Bluetooth LE beacon.… https://t.co/pecNBCR86u”

(tags: bluetooth spam via:twitter alibaba eddystone ads)

Productionproofing EKS

‘We recently migrated SaleMove infrastructure from self-managed Kubernetes clusters running on AWS to using Amazon Elastic Container Service for Kubernetes (EKS). There were many surprises along the way to getting our EKS setup ready for production. This post covers some of these gotchas (others may already be fixed or are not likely to be relevant for a larger crowd) and is meant to be used as a reference when thinking of running EKS in production.’

(tags: eks aws docker kubernetes k8s ops prod)

Neoliberalism has conned us into fighting climate change as individuals | Martin Lukacs | Environment | The Guardian

These pervasive exhortations to individual action — in corporate ads, school textbooks, and the campaigns of mainstream environmental groups, especially in the west — seem as natural as the air we breathe. But we could hardly be worse-served. While we busy ourselves greening our personal lives, fossil fuel corporations are rendering these efforts irrelevant. The breakdown of carbon emissions since 1988? A hundred companies alone are responsible for an astonishing 71%. You tinker with those pens or that panel; they go on torching the planet. The freedom of these corporations to pollute – and the fixation on a feeble lifestyle response – is no accident. It is the result of an ideological war, waged over the last 40 years, against the possibility of collective action. Devastatingly successful, it is not too late to reverse it. The political project of neoliberalism, brought to ascendence by Thatcher and Reagan, has pursued two principal objectives. The first has been to dismantle any barriers to the exercise of unaccountable private power. The second had been to erect them to the exercise of any democratic public will. […] At the very moment when climate change demands an unprecedented collective public response, neoliberal ideology stands in the way. Which is why, if we want to bring down emissions fast, we will need to overcome all of its free-market mantras.

(tags: politics environment neoliberalism future climate-change green)

Block Advertising on your Network with Pi-hole and Raspberry Pi

A good walkthrough of the Pi-Hole network-wide adblocker install and operation

(tags: pi-hole ads blocking ops home raspberry-pi)

October 21 post-incident analysis | The GitHub Blog

A network outage caused a split-brain scenario, and their failover system allowed writes to occur in both regional databases. Once the outage was repaired it was impossible to reconcile writes in an automated fashion as a result. Embarrassingly, this exact scenario was called out in their previous blog post about their Raft-based failover system at https://githubengineering.com/mysql-high-availability-at-github/ — “In a data center isolation scenario, and assuming a master is in the isolated DC, apps in that DC are still able to write to the master. This may result in state inconsistency once network is brought back up. We are working to mitigate this split-brain by implementing a reliable STONITH from within the very isolated DC. As before, some time will pass before bringing down the master, and there could be a short period of split-brain. The operational cost of avoiding split-brains altogether is very high.” Failover is hard.

(tags: github fail outages failover replication consensus ops)

LGTM

Analyses open source code repos on github, running static code analysis to find “zero-days and other critical bugs”.

(tags: security code-review static-code-analysis coding lgtm github)

cool twitter image compression hack

D?v?d ?uc??n?n on Twitter:

Assuming this all works out, the image in this tweet is also a valid ZIP archive, containing a multipart RAR archive, containing the complete works of Shakespeare. This technique also survives twitter’s thumbnailer :P

(tags: via:johnke cool compression image jpeg twitter)

A Dark Consensus About Screens and Kids Begins to Emerge in Silicon Valley – The New York Times

John Lilly, a Silicon Valley-based venture capitalist with Greylock Partners and the former C.E.O. of Mozilla, said he tries to help his 13-year-old son understand that he is being manipulated by those who built the technology. “I try to tell him somebody wrote code to make you feel this way — I’m trying to help him understand how things are made, the values that are going into things and what people are doing to create that feeling,” Mr. Lilly said. “And he’s like, ‘I just want to spend my 20 bucks to get my Fortnite skins.’”

(tags: kids technology education parenting screentime apps tech phones)

Your Kid’s Apps Are Crammed With Ads – The New York Times

In apps marketed for children 5 and under in the Google Play store, there were pop-up ads with disturbing imagery. There were ads that no child could reasonably be expected to close out of, and which, when triggered, would send a player into more ads. Dancing treasure chests would give young players points for watching video ads, potentially endlessly. The vast majority of ads were not marked at all. Characters in children’s games gently pressured the kids to make purchases, a practice known as host-selling, banned in children’s TV programs in 1974 by the Federal Trade Commission. At other times an onscreen character would cry if the child did not buy something. “The first word that comes to mind is furious,” said Dr. Radesky, an assistant professor of developmental behavioral pediatrics at the University of Michigan Medical School. “I’m a researcher. I want to stay objective. We started this study really just trying to look at distraction. My frustrated response is about all the surprising, potentially deceptive stuff we found.”

(tags: children kids ads advertising apps android google ftc games iap)

Epic twitter thread from @colmmacc explaining why client certs and mutual-auth TLS are TERRIBAD

Ok. tweet thread time! Too long ago I promised to write a screed explaining how much I hated mutual-auth TLS and why. I got distracted, and I wasn’t happy with the writing, so here it is in tweet thread form instead! But basically: Client certs and Mutual-Auth TLS is TERRIBAD. When I say TERRIBAD, I mean that unless you’ve got the resources of a big security dept and folks who comb threat models for a living, using clients certs and mutual auth probably materially lessens your security. That’s NUTS!

(tags: terribad rants twitter threads tls ssl authentication mtls security)

Google Home (in)Security

0 authentication on some setup APIs, including ‘delete the current wifi network config’ — pretty major lack of security

(tags: auth security fail google google-home)

ChromeCast HTTP APIs

similar to the Google Home hackable APIs, it seems Chromecasts are easily hacked/scripted with no auth

(tags: auth http apis security chromecast google)

“The first AI portrait in Christie’s” was mostly output from someone else’s open-source code

The print was created by Obvious, a trio of 25-year-old French students whose goal is to “explain and democratize” AI through art. Over the past year, they’ve made a series of portraits depicting members of the fictional Belamy family, amplifying their work through attention-grabbing press releases. But insiders say the code used to generate these prints is mostly the work of another artist and programmer: 19-year-old Robbie Barrat, a recent high school graduate who shared his algorithms online via an open-source license. The members of Obvious don’t deny that they borrowed substantially from Barrat’s code, but until recently, they didn’t publicize that fact either. This has created unease for some members of the AI art community, which is open and collaborative and taking its first steps into mainstream attention.[…] Jason Bailey, a digital art blogger who runs the site Artnome, says that what Obvious has done is far from unusual. “It’s almost weekly in digital art that someone takes some open code and tweaks it and sells it,” he tells The Verge. But the prominence of this auction and the fact that Obvious, not Barrat, has received the attendant prestige and attention does complicate the matter. “There’s a lot of stuff you can do that’s legal, but that makes you sort of a jerk,” adds Bailey. “If I was Robbie, I’d be pretty miffed, and Obvious said they owe him a great deal of credit.” Barrat says he holds no grudges at all and is mostly annoyed that the auction might give outsiders the wrong impression about AI art. “I’m more concerned about the fact that actual artists using AI are being deprived of the spotlight,” he says. “It’s a very bad first impression for the field to have.”

(tags: ai art graphics history open-source ownership copyright obvious robbie-barrat digital)

pusher/k8s-spot-rescheduler

‘Tries to move K8s Pods from on-demand to spot instances’:

K8s Spot rescheduler is a tool that tries to reduce load on a set of Kubernetes nodes. It was designed with the purpose of moving Pods scheduled on AWS on-demand instances to AWS spot instances to allow the on-demand instances to be safely scaled down (By the Cluster Autoscaler). In reality the rescheduler can be used to remove load from any group of nodes onto a different group of nodes. They just need to be labelled appropriately. For example, it could also be used to allow controller nodes to take up slack while new nodes are being scaled up, and then rescheduling those pods when the new capacity becomes available, thus reducing the load on the controllers once again.

(tags: k8s kubernetes aws scaling spot-instances ops)

LiV Pi

Air quality sensor board for Raspberry Pis, with a good quality self-calibrating NDIR CO2 sensor

(tags: co2 air quality monitoring metrics health home raspberry-pi hardware to-get)

Motorola and iFixit—A Match Made in Mobile

This is awesome.

Motorola is setting an example for major manufacturers to embrace a more open attitude towards repair. If you’re a Motorola customer, you can now either send in your broken device directly to Motorola for repair—or you can fix it yourself with the highest quality parts and tools, plus a free step-by-step guide, all included in our official Motorola OEM Fix Kits.

(tags: motorola repair ifixit hardware mobile)

Excellent Twitter thread from colmmacc on how s2n avoids protocol-state errors

using a linearized set of state transitions, and Cryptol and SAW to perform verification of the TLS state machine

(tags: cryptol saw formal-verification twitter threads colmmacc security s2n)

Remapping the silly double-S key on a Macbook running 10.12 Sierra

the latest hacky workaround for this stupid Apple keyboard layout bullshit

(tags: apple keyboards europe fail hidutil macosx)