875 karma · joined January 23, 2012
That begs the question, how did at least four such cutting-edge research works with fabricated data and results go through the review process unscathed?
I remember seeing several pro-China op-eds and advertorials in the NYT during the pre-Covid days, scratching my head wondering if I was in a different universe.
At the same time, there were several dozen covert and overt editorials that bashed countries and governments that were not pro-China. For example they continue to publish several blatant anti-Indian and anti-Hindu pieces which sound more like sensationalist attention-grabbing propaganda.
As a former paid subscriber to the NYT, it's sobering and sad to see how they have pivoted and where they're headed.
My MacBook pro only has USB-C ports, they work well for data, video out, as well as power (charging) and I am glad to be rid of the USB-A ports.
Turns out that we had a major outage at a data centre that served DBS - one of the largest lenders, as well as Citi. [1]
The disruption was attributed to a cooling system failure at a data centre operated by Equinix [2]. Further digging led to information that the culprit was the SG3 data centre [3] marketed as their largest IBX data centre in the Asia Pacific region and one of the newest. It turned out that the cooling system was being upgraded on contract with an external vendor, who applied incorrect settings which brought it down.
Further, this particular data centre has 2N electrical redundancy but a cooling redundancy of only N+1 chillers, in comparison to other financial services organizations like the Singapore exchange (SGX) that offers [4] CoLo hosting with 2N chillers, which I believe is essential for warm equatorial climates.
Sadly, this outage was followed by yet another smaller payment related outage the following week, making it the fifth outage this year. DBS was trumpeting their move to the cloud [5] as part of their grand plan to transform themselves from a bank into a software company that also offered banking services [6][7].
In going all out with this questionable and misguided transformation they've lost focus on what made people trust them in the first place - the decades of trust that was built on solid, reliable, transparent and efficient banking services.
There were questions about why their backup data centre didn't kick in and there are no answers till date.
It's clear to see that the recovery mechanisms weren't tested, performance degrade modes were not implemented or not tested, disaster resilience utterly failed, there were no working mitigations for cooling system failures or DC failures, and as a result ATMs and other services were down from 3pm on Oct 14 until the following morning. For a country that prides itself on digital transformation, this is just the latest banking systems failure which makes it far more than an egg in the face, it's an erosion of trust.
Items (2), (3), (7), (8), (9) from the Google report directly apply to this failure. I can only hope something good comes out of it and lessons are learnt.
[1] https://www.channelnewsasia.com/singapore/dbs-citibank-outag...
[2] https://www.zdnet.com/article/equinixs-data-center-system-up...
[3] https://www.equinix.com/data-centers/asia-pacific-colocation...
[4] https://www.sgx.com/data-connectivity/co-location
[5] https://www.dbs.com/newsroom/First_bank_in_Singapore_to_laun...
[6] https://bankinginnovation.qorusglobal.com/content/articles/t...
[7] https://www.dbs.com/technology-future/dbs-redefining-the-fut...
USB is by far the worst designed thing in the history of humankind, in terms of the cumulative amount of frustration it has caused globally.
All common-sense usability principles were ignored, and these are not new principles. Take the old RS-232C or RJ45 jacks, you just looked at it and you knew which way to plug it in. Hell you could do it blindfolded by feel.
There are several other passages from the oldest strata of Pāli texts, e.g., the SNP (Sutta Nipata) and some of the connected discourses (Samyutta Nikāya) that talk more about metta, a rich and complex tapestry that paints a picture of the actual and original intent behind the word or phrase.
This talk by Prof. John Peacock has some good insights into the phrase mettā, and provides a good overall context.
https://www.audiodharma.org/talks/2600
Mettā is part of the four brahmavihārās or practises to cultivate wholesome states of mind, the others being karuṇā (compassion & kindness for oneself and for others), muditā (simplistically translated as "empathetic joy") and upekkhā (equanimity).
China launches a $40 billion state fund to boost the chip industry.
https://www.reuters.com/technology/china-launch-new-40-bln-s...
Archive link: https://archive.ph/1lWyv
I've had the pleasure of deep-dive analysis and problem solving with some extremely talented and intelligent minds there. Though we had a language barrier, it became irrelevant after a few discussions.
What also amazed me about a company of that size was their decision making process and ability to commit resources and money to actions at very short notice. Their divisions and teams operate in a very autonomous style, and the teams I interacted with were definitely empowered to make strategic calls and commit to them. This may not be true across the whole organisation, but it does broadly appear to be an indicator of their culture.
Back then, they were already doing a fair amount of work on upstream cellular networking and related tech around 5G, SOC design, manufacturing processes and improvements, and a whole lot more. They were not shy about discussing sanctions and were (still are) hard at work towards self-sufficiency, sanctions or not.
What's happening now was not only expected, it could not have turned out otherwise. Also note, we're hearing about Huawei because they are one of the biggest. There are several other smaller and niche players who are part of the game.
Aside from AI-generated garbage, there are large numbers of works that are of just poor quality, with little to no editing, spelling mistakes, grammatical mistakes, formatting and layout problems, and (on one occasion) very smelly paper.
AI-generated rubbish is only the latest incarnation of this problem. I recall many years ago, purchasing a book on health and fitness that had many 4- and 5-star reviews. It was not cheap. It turned out to be a 50-60 page booklet that looked like someone had taken a bunch of webpages and clicked "print to PDF" on the browser - absolutely no proofreading or formatting. There were pages with the old IE missing image icon in place of what should have been photographs.
There is no reliable curation, a reputation scheme, or any trustworthy process on amazon to discover new works or explore new topics that return genuine material. This also extends to other goods such as household items, utility items, foods, and so on. It's scary that amazon also sells medicines - what's the guarantee that they are not fake, adulterated or toxic?
Amazon is a cesspool.
The example shows you how to use the swagger tool, parse the OpenAPI spec [2], auto-generate GoLang glue code, call __one__ of those auto-generated functions and log a trace.
However, there is zero documentation, zero other examples, and I'm left scratching my head whether there's even one person in the world using this approach. I eventually ended up just directly using the service APIs [3] via REST calls.
OTEL is painful, but the alternatives are no better :( I really wish there's some interest in this space, since SLO's and SLI measurements are becoming increasingly important.
[1] https://github.com/openzipkin/zipkin-api-example
[2] https://github.com/openzipkin/zipkin-api/blob/master/zipkin2...
Consider this scenario: There is a collection of services that talk to one another, and not all use HTTP. Say agent A0 makes a connection to agent A1, this is observed by service S0 which triggers service S1 to make calls to S2 and S3, which propagate elsewhere and return answers.
If we limit the scope of this problem to services explicitly making HTTP calls to other services, we can easily use the Propagators API [1] and use X-B3 headers [2] to propagate the trace context (trace ID, span ID, parent span ID) across this graph, from the origin through to the destination and back. This allows me to query the metrics collector (Jaeger or Zipkin) using this trace ID, look at the timestamps originating at the various services and do a T_end - T_start to determine the overall time taken by one call for a round trip across all the related services.
However, this breaks when a subset of these functions cannot propagate the B3 trace IDs for various reasons (e.g., a service is watching a specific state and acts when the state changes). I've been looking into OTEL and other related non-OTEL ways to capture metrics, but it appears there's not much research into this area though it does not seem like a unique or new problem.
Has anyone here looked at this scenario, and have you had any luck with OTEL or other mechanisms to get results?
[1] https://opentelemetry.io/docs/specs/otel/context/api-propaga...
In my organization if I were in charge of contracts and pulled something like this, I'd be out of a job and my chain of command would be under review.
This is simply unprofessional.
The telcos I deal with don't even have a business model or a way to price enterprise services delivering these new 5G features, they're just selling a pipe with newer tech and asking for more money.
There are a bunch of companies, research organisations, and universities building and making demos but hardly any of these make it to production. Here's an example of a working system: Local 5G and TETRA interconnected at Narita International Airport: https://www.ntt-east.co.jp/en/release/detail/20230519_01.htm...
I work in this space... 5G coverage is a lot worse in most places than 4G or even 3G. As for the overhyped applications like remote surgery, remote tele-operated robots, 5G-driven cars, etc., don't hold your breath.
mmWave 5G is a dud - it has some promise but again telcos are not investing money to roll out mmWave. Telcos aggressively marketed 5G to consumers at a premium price point but they offer very limited benefits in limited situations. Even routine functions like cell to cell handoffs are often smoother in 4G than with 5G, not because of the spec but because of other factors.
5G was overhyped and it failed to deliver any value except in a few situations.
Though he talks about the five properties, I tend to disagree with this generalisation - what I like at the moment really depends on my state of mind, mood, time of the day, and what I want to feel like after listening to a few songs. Sometimes I really get into songs with complex patterns and interesting time-signature changes, and at other times I simply settle with some Wes Montgomery. What sounded good to me this morning doesn't work for me now. What I disliked five years ago is something I keep coming back to today.
Age contributes as a very significant factor to what sounds good. Speed metal or power metal was "good" in my teens and twenties, but these days I need to be in a very specific mood with specific energy levels to even consider some Pantera or Megadeth. On the other hand, I can easily get into Sepultura ("Ratamahatta", "Roots bloody roots", etc.) even though they are often much heavier, because there's a kind of easy coexistence with their type of music that I find curious. There are other musicians I liked way back then, but you couldn't pay me money to listen to them today - even the same songs I had on repeat.
Perhaps the prof is on theoretically sound footing, I'm a layman in this field, but I can't experience what he describes.
Would it work like the no-code platforms where you're 100% dependent on those providers to even make your software run?
I only see red flags for corporate use. Sure, this may be nice to play with a few ideas, but I would definitely not trust it to host and build production software.
How do you get a whole organization of several thousand people to move from zoom to another platform? In my experience, this has only happened because the provider simply could not meet the QoS/ToS, or they were priced out.
Also, most people just don't care. They just want something working well. We've got Cisco webex and Microsoft teams, but guess what most people prefer? Zoom. Because "it just works better".
Take google meet for instance. There are already two versions of this app on the play store. One is "google meet (original)" and the other is just "google meet". The weird thing is that both of them are regularly updated, and the "(original)" app was last updated just a week ago.
As for alternatives like Jitsi, perhaps I can get it going, and perhaps a few people would join for a few weeks. The interest will slowly die out as people go back to what they know - zoom or teams. My mother in her 70's uses zoom to collaborate with others on some work that she does - they understand the privacy issues but they simply won't accept the hassle of changing to something else. Not unless zoom does something really bad.
The only reason I was using PayPal is so that I didn't have to provide my credit card number to merchants while purchasing online. Now with single-use CC numbers, I haven't used PayPal for a few years.
What am I missing?
And another $100K to an "Action research collaborative" that sounds like another social justice org?
WTF? This sounds like money laundering. I just want a good browser. I'm now regretful that I was donating regularly to Mozilla.
Imposing such an artificial and rigid structure when attending a talk will stress me out, get me less, and turn me off. I still remember my student days when the best lectures or talks served to inspire, and I'd often dig deep into a subject because of that.
Be specially careful when installing their extension pack, as it is an evaluation license.
We've moved our development to KVM and Virtual Machine Manager on Linux [3] and UTM on Mac [4]. There are other options to run your VM, such as Multipass [5] or VirtualBuddy [6].
On a digressive topic - it was fun migrating our legacy application server stack from Oracle Java (old & poorly considered decision) to OpenJDK, thanks to their license [2].
[1] https://www.virtualbox.org/wiki/Licensing_FAQ
[2] https://www.oracle.com/java/technologies/javase/jdk-faqs.htm...
Migration of my mails from Mozilla to thunderbird was trivial since they used the same mbox format. I've moved my thunderbird profile after that from Linux to Windows to Mac, across nearly 20 years, with no problems.
I will continue to use Thunderbird for the simple reason that it just works on all the common OS'es and it's easy to move your profile across them.
This is yet another large and powerful company acting to appease the market so their stock price goes up a bit and the shareholders are rubbing their palms in glee.
Now I'm off to watch "Office Space"
They are already saving a significant amount of money by giving up our old offices. The mode of engagement between staff and across teams has obviously changed and continues to do so, but productivity is back to pre-Covid levels. We meet in person when there's a real need, say once a week or two.
Even allowing for the possibility that the contents of the article are true (which I seriously doubt ... it reads like sloppy reporting and cheap sensationalism) this is one man's view, biased and coloured by his own interests.
So may great hits are reinterpretations of older less-known numbers, often as tributes to the musicians who wrote or played the originals.
What if the estates of Bobby Russell won against Reba McEntire, or the estates of Sonny Boy Williamson or Willie Dixon won against Led Zeppelin, heck, pick pretty much any old blues musician and you can claim they've been ripped off by newer bands. Freddie King's estate would be rolling in dough and their heirs would be set forever.
The problem lies with copyright laws, the way they are today. There will always be greedy money-grubbing lawyers and people wanting to make easy money by exploiting this system. This goes beyond copyright laws. Patent laws anyone? What if Volvo had decided to charge an arm and a leg for their 3-point seatbelt patent instead of letting competitors use the idea without charge? Nothing wrong about that, they may have made a ton of money (justifiably so), but they looked at the broader picture concerning public safety. Would Chrysler have done that had they invented the 3-point seatbelt? Probably not. I'm just trying to illustrate how two parties could make completely opposite decisions both compliant with copyright or patent law.
The example in the second link shows the details that the cellular network provider may have, and this includes the Cell tower ID, the enodeBID (4G), tracking area (TA) ID, geolocation, information about whether the device is moving (its velocity), the mobile country code (MCC), mobile network code (MNC), and a whole lot more.
This data is already with your cellular network provider, so it makes no difference if you have "switched off" location services on your phone.
[1] The Open5G Lab APIs https://swagger.open5glab.net/
[2] An example location tracking request https://swagger.open5glab.net/#/MonitoringEvent%20API%20Subs...
My message was then promptly sent to the very same hiring manager running the show, and the local director who was equally involved. Obviously, nothing positive came out of it, and there was a fair amount of unpleasantness afterwards.
Back then I was younger and somewhat naïve, but fortunately that mindset changed quickly. None of these hotlines or whistleblower mailboxes work as intended. More often than not, you just end up creating trouble for yourself and for others with a net negative outcome while the original problem remains unaddressed.
The people in the positions of power have everything going for them, it's in their interests to maintain status-quo and expand their spheres of influence. The last thing they want is for an upstart whistleblower to try and make a change.