Tag Archives: OpenAI

Major Outages in 2024

2024 was like most recent years where there were a few major broadband outages and a lot of smaller regional ones. Most carriers claim to be investing more money in increased redundancy to avoid major outages and one hopes that is cutting down on outages.

AT&T suffered a big outage in February when it lost cellular coverage in markets like Dallas Houston, Los Angeles, and Atlanta. The outage particularly affected first responders served by AT&T’s FirstNet network. The company said the outage was “caused by the application and execution of an incorrect process used as we were expanding our network, not a cyberattack” Basically, the company messed something up during a network update.

The biggest telco outages for the year came from hurricanes. In western North Carolina alone, 80% of cell sites went out of service by the day after the storm hit. Fiber networks were severed as entire roads washed away, and something like a million trees were damaged. I live in Asheville, and we experienced a total communications blackout with no cellular or landline broadband. It took about a week to get a partial cell signal back and over three weeks to get broadband. Some rural areas were out much longer.

Hurricane Milton caused broadband outages as well, more related to power outages than destroyed telecom network. A lot of places didn’t lose cell coverage, and most people were back in service within a few days.

The other big outages in 2024 were not network outages but service provider outages.

  • Microsoft Teams had a seven hour outage on January 26. The cause of the outage was never disclosed but seems to have been internal to Microsoft.
  • On March 5, Meta had an outage that blocked users from accessing Facebook, Instagram, Messenger, and Threads. The reason for the outage was a glitch in the login process.
  • Google lost service for an hour on May 1. The problem was a failure in the verification process that couldn’t identify users.
  • The biggest outage of the year happened on July 19 and affected 8.5 million Microsoft Windows devices. The outage was worldwide. Flights were canceled, customers couldn’t access banks, surgeries were canceled, and there were widespread 911 outages. The cause of the problem was a section of code at CrowdStrike, the cybersecurity firm that many large Windows customers were using to protect their devices. In retrospect, the outage was blamed on the lack of testing from CrowdStrike before implementing a software update.
  • Microsoft had an outage on November 25 that caused intermittent inability for users to use Outlook or reach the web. Microsoft admitted the source of the problem was a configuration change – another software update problem.
  • On December 11, OpenAI had an outage of it’s video service Sora. This was caused by a cascading error when a telemetry service overwhelmed the platform.

Interestingly, most of the service outages were the result of configuration changes, meaning software upgrades.

These big companies should learn a lesson from smaller telcos. I’ve had many clients who learned the hard way to never introduce new software onto customers without first testing the update. That means NEVER EVER, NEVER EVER, NEVER EVER (did I say that enough?). Many telcos have software test labs where they have a lab setup that mimics the network. They try updates in the test lab before ever subjecting their customers to an untested update. This is software update 101 stuff, but apparently, the smart guys at some of the biggest companies don’t think they need to take this extra precaution.

Competing with ChatGPT

I’ve been writing this blog daily since 2013, and writing it is the favorite part of my day. Writing the blog forces me to research and solidify my thoughts and opinions about various topics. But suddenly, I’m seeing headlines everywhere saying that ChatGPT will soon handle most writing and there will be no need for folks like me who write every day.

I was obviously intrigued and investigated the ChatGPT software. The latest 3.5 version of the software was launched by OpenAI in November 2022. OpenAI is a for-profit software firm that has been researching the field of artificial intelligence (AI) with the stated goal of developing friendly AI. It’s interesting that friendly is a key part of their mission statement because many AI industry pundits predict that AI will likely eventually compete with humans for resources, much like Skynet in the Terminator movies.

ChatGPT is written atop OpenAI’s third generation of software and is aimed at communicating in a written or conversational way so that a reader can’t tell the difference between the software and a human. The company has numerous investors, but Microsoft just offered to buy a 49% stake in the company for $10 billion. This instantly has me wondering when there will be a fee to use the software instead of the free version that is available now.

The press on ChatGPT has been over-the-top. I’ve seen articles comparing the impact of the launch of ChatGPT to other big events in web history, like the first web browser or the iPhone. Articles are touting that the software will mean that programmers will no longer have to write code, that students will no longer have to write papers, and that there will soon be no need for journalists (or bloggers!)

Early-generation AI writing software has been around for a few years and many baseball box scores and press summaries of quarterly earnings reports have been generated by software. These are writing tasks that are formulaic and repetitive, and I doubt that most folks noticed – although the software never captured the magic of a sports reporter like Shirley Povich, who I enjoyed reading every day for years in the Washington Post.

I had to give this platform a try. Was this software capable of writing something like this blog? If so, it would make me reconsider writing every day because if the software is that good there won’t be much need for human writers before long. As I was testing, I also considered the idea of using the software to get a jump start on a new piece of writing – the idea of seeing if the software could structure and organize an idea would be a time saver if the results were usable.

You can give complicated instructions to the software. You can provide the topic, the desired length of the end product, and describe the desired style of writing. I gave the software several topics to write about, and I was impressed with the speed of the process. The finished product is created almost as soon as you say go to the software.

But I was underwhelmed by the results. The sentences are grammatically perfect, and each paragraph has a topic and tries to make a point. Yet the end result was stilted, and some paragraphs were unreadable – I had to reread them several times to try to decipher the point (but for all I know, my readers have to do the same thing!).

The biggest flaw was that the writing was full of factual errors. That makes a lot of sense because the software distills what is written on the web when writing content. It takes the good and the bad, the factual and non-factual, and the easy-to-understand and obtuse writing that exists on the web and mashes it in a synthesis of what it finds. I realized that I would have to fact-check everything ChatGPT writes because the software has no way to discern what is true or untrue. There is a term for this among data scientists, and I read that ChatGPT currently has a hallucination rate of between 15% and 21%, meaning that it seems to make up that percentage of facts in its writing.

I know there is instant hope among students that this software can churn out the dreaded school essay – but that doesn’t look likely. The software has been out for only two months, and I saw that a software engineer has already developed a program that can detect with more than 90% accuracy if something is written by a human or by ChatGPT. Students beware.

The day will likely come when the ChatGPT writing gets better, but there is nothing in this software today that would make me consider giving up writing or even using this as a tool. The hallucination rate means I can’t trust it to be factual, so it’s not even worth using to create a kernel of an idea for a blog. Most importantly, the output is not readable – it’s all perfect English, but I couldn’t understand the point of about half of what it wrote for me. If my blogs are going to be unreadable, I want the obtuseness to be fully human-generated!

New Technologies, June 2017

Following are some interesting new technologies I’ve run across recently.

WiFi Imaging. Cognitive Systems has a product they call Aura that can detect motion inside of a home using WiFi. The technology was developed a few years ago at MIT. The technology used is called Radio Frequency (RF) Capture. The device can sense subtle changes in wireless signals to determine if something is moving in the home. It can be set to different sensitivities to be able to detect people, but not animals. It can also be set to track specific cellphones so that you’ll know when a known person has entered or left the home. For now the device does not connect to external security services but sends a message to a smartphone.

Some German researchers at the University of Munich have already taken this same idea a lot farther. In a paper published in the Physical Review of Letters they describe a technique where they can use WiFi to create 3D holographic images through walls. The lab unit they have built can detect objects down to about 4 centimeters in size. They scan ten times per second and can see outlines of people or pets moving inside of another room. This technology is eerily reminiscent of the surveillance machine in The Dark Knight that Bruce Wayne destroys at the end of the movie since it was a scary invasion of privacy.

Eliminating IoT Batteries. One of the scariest things about the exploding number of devices used for IoT is the need to power them, and the potential huge waste, cost and hassle of needing batteries for tons of devices. Tryst Energy from the Netherlands has developed an extremely efficient solar device that only needs 200 lux of light for four hours per day to operate a small sensor that communicates with Bluetooth or WiFi. That is the amount of light normally found under a desk. The device also ought to last for 75 – 100 years, opening the ability to place small IoT sensors in all sorts of places to monitor things. When you consider the expected billions of devices that are expected over the next decade this could provide a huge boost to the IoT industry and also provide a green solution for powering tiny devices. The device is just starting to go into production.

Bots Have Created Their Own Language. A team at OpenAI, the artificial intelligence lab founded by Elon Musk and Sam Altman, has published a paper describing how bots have created their own language to communicate with each other. They accomplished this by presenting simple challenges that require collaboration to bots, which are computer programs that are taught to accomplish tasks. Bots are mostly being used these days to learn to communicate with people. But the OpenAI team instead challenged the bots to solve spatial challenges such as devising a way to move together to a specific location inside of a simulated world. Rather than tell the bots how to accomplish this they simply required that the bots collaborate with other bots to accomplish the assigned tasks. What they found was that the bots created their own ‘language’ to communicate with each other and that the language got more efficient over time. This starts sounding a bit like bad Sci-Fi world where computers can talk to each in languages we can’t decipher.

Recycling CO2. Liang-shi Li at Indiana University has found a way to recycle CO2 for the production of power. He has created a molecule that, with the addition of sunlight, can turn CO2 from the atmosphere into carbon monoxide. The carbon monoxide can then be burnt to create power, with the byproduct being CO2. If scaled this would provide for a method to produce power that would add no net CO2 to the atmosphere (since it recycles the CO2). Li uses a nanographene molecule that has a dark color and that absorbs large amounts of sunlight. The molecule also includes rhenium which is then used as a catalyst to turn nearby CO2 into carbon dioxide. He’s hoping to be able to accomplish this instead with more easily obtained magnesium.

Liquid Light. It’s common knowledge that light usually acts like a wave, expanding outward until it’s reflected or absorbed by an object. But in recent years scientists have also discovered that under extreme conditions near absolute zero that light can also act like a liquid and flow around objects and join back together on the other side. The materials and processes used to produce the liquid light are referred to as Einstein-Bose condensates.

Scientists from CNR Nanotec in Italy, Ecole Polytechnique de Montreal in Canada, and Aalto University in Finland just published an article in Nature Physics that shows that light can also exist in a ‘superliquid’ state where light flows around objects with no friction. Of most interest is that this phenomenon can be produced at normal room temperature and air pressure. The scientists created this effect by sandwiching organic molecules between two highly-reflective mirrors. The scientists believe that interaction of light with the molecules induces the photons in the light to take on characteristics of electrons in the molecules.

The potential uses for this technique, if perfected, are huge. It would mean that light could be made to pass through computer chips with no friction, meaning no creation of the heat that is the bane of data centers.