Bing Gaffes: Errors Made by the Softies

February 16, 2023

Microsoft does a bang up job of marketing. I am not sure the security of the firm’s systems and the ability to allow users to print are comparable. But Microsoft has mindshare.

train 1 to use

The images of commuters who missed their search express train are the brilliant work of Craiyon.com smart software. This write up is the work of the dinobaby Stephen E Arnold.

The word “google” now evokes pushback. MSFT’s mission is accomplished. To remind people of the missteps in smart Bing, Simon Willison has compiled an interesting list of flubs. These range from errors about vacuum cleaners to a few threats to users. If you enjoy looking at what happens when smart software demonstrates the thrill of recursive learning, you will want to read “Bing: “I Will Not Harm You Unless You Harm Me First”.  I quite liked the essay. I am confident that the Google smart software team will distribute the examples to the wizardly Googlers. Forewarned of embedded and poorly understood error generation is a good thing. Too bad neither Google nor Microsoft considered the issue before the PR tsunami and the new arms race.

Stephen  E Arnold, February 16, 2023

The Microsoft Vision: Agent-Intermediated Computing

February 9, 2023

Not one single input from smart software. — Editor

[Mise en scène] Googzilla is towering over advertisers, hissing and brandishing its long talons. Then the creature turns its head. It tiny ear slits twitch. He hears a sound like “boink” or “thunk.” Distracted, the 25 year-old terror looks from the cowering advertisers and fixes his maroon-hued eyes on a almost insignificant figure. That entity is Satya Nadella, who has just ruing Sundar’s and Prabhakar’s high school reunion. The fear of answering the question “Hey, how did you guys miss this ChatGPT – Microsoft thing” is terrifying. Googzilla emits a plaintive “welp.” The advertisers back away and start walking toward Redmond. [Fade to black]

The shoe has dropped. Boink or thunk, depending on your perceptual equipment. “Microsoft’s AI-Powered Bing Will Challenge Google Search” reports “Microsoft may finally have figured out how to get you to use Bing.” The article adds a quote directly from the champion of high school reunions in India:

“All computer interaction is going to be mediated with an agent helping,” Chief Executive Satya Nadella said at a launch event at the company’s headquarters in Redmond, Washington. “We’re going to have this notion of a co-pilot that’s going to be there across every application.”

I won’t point out that “all” is a logical impossibility when it comes to humanoids and computer systems. But I will let that slide… because ChatGPT.

Kudos to Microsoft for pulling off the marketing play of the year. I know it is only early February 2023, but it may take something truly special to tap into the brush fires ignited by ChatGPT. Will Bing be better? Maybe? Will the ChatGPT thing frighten the allegedly monopolistic Google? I think it has.

There are several examples that illustrate the disarray of the Google radar system. First, the search beacon missed the incoming ChatGPT balloon. Hello, Prabhakar. Isn’t and maybe wasn’t that your job?

Then there was the startling Code Red. Yeah, that’s professional. OpenAI has been around six or seven years. Now it is Code Red. Situational awareness seems to be lacking where Googzilla hunts. This is a flaw in an apex predator, is it not?

The dollop of whipped cream on this torched cupcake was asking the former head chefs of data hoovering and search engine optimization as an spur to buying advertising to return to the wizard lair. Yep, Sundar asked Sergey and Larry to help out with the Code Red thing. Okay, but let’s recall the origin of the Google money machine. Wasn’t it Yahoo-Overture-GoTo? Yeah, I think it was and there was a legal hassle and a billion dollar sentiment to make the GOOG gleam like a sprightly young Googzilla.

The actions that cement the frenzy in the House of Google is the steady flow of “it’s coming,” “yes, it’s a demo,” and “okay, we bought an outfit that Sam Bankman Fried found interesting.” The problem is that “to be” does not close the gap with the ChatGPT riding its hyper-drive electronic motorcycle on Google’s private motorways.

Several observations:

  1. Will the OpenAI and ChatGPT thing help Microsoft address the security of its existing software and services? When?
  2. Will Microsoft milk the marketing buzz and return to business as usual: Killing printers, interfaces that baffle, and features that disrupt one’s activity on a Windows enabled system?
  3. Will Microsoft have an answer to those who would claim that smart software violates fair use of intellectual property?
  4. Will Microsoft be able to handle bias and outputs which lead to interesting but harmful outcomes like a student getting expelled after mommy and daddy paid $135,000 for tuition?

But for now. The payoff for Microsoft is the thrill of watching Googzilla squirm. And the “all” word? That’s just an illustration of the imprecision of Microsoftie speech.

Stephen E Arnold, February 9, 2023

Google and OpenAI: The Big Dust Up

February 8, 2023

Let’s go back to high school English class and a very demanding spinster named Miss Drake. Her test question was, “Who wrote this line?”

O heaven! that one might read the book of fate, and see the revolution of the times. (Henry IV, Part 2 [~ 1597] Act 3)

Do you remember? I do. The answer is the Bard of Avon, and he allegedly spun the phrase from his brain cells or he ripped it off from another playwright. Yeah, the plagiarism thing has not been resolved, and it is unclear if and with whom the Bard sucked in content and output a money-making play. Was the real Bard a primordial version of a creator on YouTube? Sure, why not draw that connection?

Now back to the quote. I like the idea of “the revolution of the times.”

The idea is that some wizard like Sundar or Prabhakar can check out the “Book of Fate” which may or may be among the works in the Google data warehouse and see the future. Just the Palantir seeing stone which works so darned well as those SPAC bets attest. Perhaps that’s what happened when Google declared a Code Red? Fear and a big bet that the GOOG can de-momentum ChatGPT.

When did OpenAI become a thing? I would suggest that it was in 2015 if one believes Sillycon Valley history. The next highlight of what was something of note took place in 2019 but possibly earlier when Microsoft plopped some Azure cycles on the OpenAI conference table. Two years later we get this:

Google at Code Red over ChatGPT As Teams Reassigned to Work on Competing AI Products

Almost coincident with Google’s realizing that people were genuinely excited about ChatGPT, Google realized that Prabhakar’s hair had caught on fire. Sundar, the Sillycon Valley manager par excellence called back the original Relevance Revolutionaries (Messrs. Brin and Page) after Microsoft made it evident to the well-fed at Davos that Softies were good at marketing. Maybe Microsoft fell short of the “Start Me Up” for the outstandingly “good enough” Windows 95, but the Chat GPT deal is notable. To make sure the GOOG got the message that Microsoft was surfing on the waves created by ChatGPT, another bundle of billions were allocated to OpenAI and ChatGPT. The time was January 2023, and it was clear that millions of norms interested in Microsoft’s use of ChatGPT in those “good enough” engineering marvels, Bing.com and the Google infused Edge browser.

Where are we on Wednesday, February 8, 2023? How about this as a marker:

Google said Bard would be widely available to the public in the next few weeks. Source: MSN.com

Yep, the killer words are right there—”would be.” Not here, the conditional future, just a limited test. Not integrated into heaven knows how many apps like OpenAI. Not here like the collection of links generated by Matt Shumer. Not here like the YouTube videos explaining how to build an app from scratch with ChatGPT. Nope. Google is into the “to be” and “demo” mode.

Let’s do simple math on Google’s situational awareness:

  • 2015 OpenAI and Elon Musk are visible
  • 2019 Microsoft provides some “money” which may be a round trip to pay for Azure cycles
  • 2022 (November) ChatGPT gets a million users in five days
  • 2022 (December) Google feels the heat from burning hair and screeches “Code Red”
  • 2023 (January) Davos talks about ChatGPT, not just the new world order, power, money, and where to eat dinner
  • 2023 (February) Google says, “Out version of ChatGPT is coming… soon. Really, very soon.” But for now it’s a limited demo. And Microsoft? The ChatGPT thing turned up when one of my team ran a query on Tuesday, February 7, 2023. Yep, ready or not there it was.

Several observations:

  1. The Sundar and Prabhakar duo missed the boat on the impact of ChatGPT for search. Err, you are supposed to be the search wizards, and you are standing on the platform waiting for the next train to arrive?
  2. The uptake of ChatGPT may be a reaction against the Google search system, not a reaction to the inherent greatness of ChatGPT. A certain search company with a 90 percent share and a genuine disdain to relevant responses to users’ queries may have boosted the interest in ChatGPT. If I am correct, this is an unintended consequence of being Googley.
  3. The Microsoft marketing move is an outstanding [a] bit of luck, [b] a super star tactic that warrants a Grammy (oh, wait, the Grammies suck), or [c] a way to breathe new life into products which suffer from featuritis and a lack of differentiation.

Net net: Sundar and Prabhakar are destined for Harvard case study glory. Also, the dynamic duo may pull a marshmallow from the fire, but will it make a great S’more? Does sophomoric rhyme with Anthropic. And Code Red? The revolution of the times might be here as the Bard wrote or obtained from a fellow playwright.

Stephen E Arnold, February 8, 2023

More on the AI Betamax Versus VHS Dust Up

February 2, 2023

24 Seriously Embarrassing Hours for AI” gathers four smart software stumbles. The examples are highly suggestive that some butchers have been putting their fingers on the scales. The examples include the stage set approach to Tesla’s self driving and OpenAI’s reliance on humans to beaver away out of sight to make outputs better. In general I agree with the points in the write up.

However, there is one statement which attracted my yellow high light pen like a sci-fi movie tractor beam. Here it is:

Sometimes the slower road is the better road.

It may be that the AI TGV has already left the station and is hurtling down the rails from Paris to Nimes. Microsoft announced that the lovable Teams video chat and Swiss Army knife of widgets will be helping users lickity split. Other infusions are almost certain to follow. Even airlines are thinking smart software. Airlines! These outfits lose luggage with bar codes. Perhaps AI will help, but I remain skeptical. How does one lose a bag with a bar code in our post 9/11 world?

The challenge for Google, Facebook (which wants to be a leader in AI), and the other organizations betting their investors’ money on AI going to take a “slower road”?

My TGV high speed train reference is not poetical; it is a reflection of the momentum of information. The OpenAI machine — with or without legerdemain — is rolling along. OpenAI has momentum. With foresight or dumb luck, Microsoft is riding along.

The “slower road” echoes Google’s conservative approach. Remember that Google sacrificed credibility in AI with the Dr. Timnit Gebru affair. Like a jockey with a high value horse, the beast is now carrying lead pads. Combine that with bureaucratic bloat and concern for ad revenues, I am not sure Google and some other outfits can become the high twitch muscled creature needed to cope with market momentum.

Betamax was better. Well, it did not dominate the market. VHS was pushed into the ditch, but that required time and technological innovation. The AI race is not over but the “slow” angle is late from the gate.

Stephen E Arnold, February 2, 2023

Microsoft Security and the Azure Cloud: Good Enough?

January 27, 2023

I don’t know anything about the cyber security firm called Silverfort. The company’s Web site makes it clear that the company’s management likes moving icons and Microsoft. Nevertheless, “Microsoft Azure-Based Kerberos Attacks Crack Open Cloud Accounts” points out some alleged vulnerabilities in what Microsoft has positioned as its present and future money machine. The article says:

Silverfort disclosed the issues to Microsoft, and while the company is aware of the weaknesses, it does not plan to fix them, because they are not “traditional” vulnerabilities, Segal says. Microsoft also confirmed that the company does not consider them vulnerabilities. “This technique is not a vulnerability, and to be used successfully a potential attacker would need elevated or administrative rights that grant access to the storage account data,” a Microsoft spokesperson tells Dark Reading [the online service publishing the report].

So a nothingburger (wow, I detest that trendy jargon). I would view Microsoft’s product with a somewhat skeptical eye. Bad actors show some fondness for Microsoft’s approach to engineering.

Shift gears, the article “Microsoft Is Beating Google at Its Own Game.” I thought, “Advertising.” The write up has a different angle:

Following the news of Microsoft’s $10 billion investment, Wedbush analyst Daniel Ives wrote that ChatGPT is a “potential game changer” for Microsoft, and that the company was “not going to repeat the same mistakes” of missing out on social and mobile that it made two decades ago. Microsoft “is clearly being aggressive on this front and not going to be left behind,” Ives wrote.

Yep, smart software. I think the idea is that using OpenAI as a springboard, Microsoft will leapfrog into high clover. The announcement of Microsoft’s investment in OpenAI provides compute resources. If the bet pays off, Microsoft will get real money.

However, what happens when Microsoft’s “good enough” engineering meets OpenAI.

You may disagree, but I think the security vulnerabilities will continue to exist. Furthermore, it is impossible to know what issues will arise when smart software begins to think for Microsoft systems and users.

Security is a cat-and-mouse game. How quickly will bad actors integrate smart software into malware? How easy will it be for smart software to trawl through technical documents looking for interesting information?

The integration of OpenAI into Microsoft systems, services, and software may require more than “good enough” engineering. Now tell me again why I cannot print after updating Windows 11? Exactly what is Google’s game? Excitement about what people believe is the next big thing is one thing. Ignoring some here-and-now issues may be another.

Stephen E Arnold, January 27, 2023

Google, Take Two Aspirin and One AlkaSeltzer: It Is Buzz Time for ChatGPT

January 17, 2023

What do you do when the “trust” outfit Thomson Reuters runs a story with this headline? “Davos 2023: CEOs Buzz about ChatGPT-Style AI at World Economic Forum.” If you are like me, you think, “Meh.”

But what if you are a Google / DeepMind wizard?

Now consider this headline: “Google’s Muse Model Could Be the Next Big Thing for Generative AI.” If you are like me, you think, “Sillycon Valley PR.”

But what if you are an OpenAI or Microsoft brainiac?

In terms of reach, I think the Reuters’ story will be diffused to a broader business audience. The subject is something perceived as magnetic. Any carpetlander can get an associate to demonstrate ChatGPT outputting a search result via You.com or some other knowledge product from the numerous demos available with a mouse click.

But to see the Google Muse story, one has to follow a small number of Sillycon Valley outlets. And what if the carpetlander wants to see a demonstration of the magical, super effective Muse? Yeah, use your imagination.

Perhaps Google and its ineffable search team may want to crunch on another couple of aspirin and get some of that chewable antacid stuff. It is going to be a long PR day at Davos.

One doesn’t have to be a business maven to understand that ChatGPT is a nice subject when the options at Davos are war, plummeting demand for some big buck commodities, Germany’s burning lignite, China’s Covid and Taiwan fixation, and similar economically interesting topics.

What will CEO and Davos attendees take away from the ChatGPT buzz? My experience suggests some sort of action, even it is nothing more than investigating whether the technology can deal with pesky customer support inquiries.

And where is Google amidst this buzz? Google has the forward forward, next big thing. Google has academic papers which point out the weaknesses of non Google methods. Google has Muse or at least a news release story about Muse.

Will OpenAI and ChatGPT have legs? Who knows. Good bad or indifferent, ChatGPT has buzz, lots of it. I know because the “trust” outfit says ChatGPT will “transform” the security minded Microsoft. Who knew?

Thus, at this moment in time, Google may become a good customer for over-the-counter headache remedies and AlkaSeltzer. Remember that jingle’s lyrics?

Plop plop, fizz fizz

Oh, what a relief it is.

Maybe ChatGPT will just fade away like hangover or the tummy ache from eating the whole thing? Is it my imagination or is Microsoft chowing down on croissants whilst explaining what ChatGPT will do for its enterprise customers?

I will consult my “muse.” Oh, sorry, not available.

Stephen E Arnold, January 17, 2023

Microsoft Reveals Its Engineering Approach: Good Enough

January 5, 2023

I was amused to read “State of the Windows: How Many Layers of UI Inconsistencies Are in Windows 11?” We have Windows 11 running on one computer in my office. The others are a lone Windows 7, four Windows 10 computers, four Mac OS machines with different odd names like  High Sierra, and two Linux installations with even quirkier names. Sigh.

The article does a masterful job of pointing out that vestiges of XP, Vista, Windows 7, and Windows 8 lurk within the Windows 11 system. I have shared my opinion that Microsoft pushed Windows 11 out to customers to deflect real journalists’ attention from the security wild fire blazing in SolarWinds. Few share my viewpoint. That’s okay. I have been around a long time, and I have witnessed some remarkable executive group think when a crisis threatens to engulf a bonus. Out she goes.

But the article makes very, very clear how Microsoft approaches the engineering of its software and systems. Think of a lousy cake baked for your 12th birthday. To hide the misshapen, mostly inedible mess, someone has layered on either Betty Croker-type frosting in a can and added healthy squirts of synthetic whipped “real” cream. “Real,” of course, means that it squeaked through the FDA review process. Good enough.

Here’s one example of the treasures within Windows 11. I quote:

The Remote Desktop Connection program is still exactly the same as it was 14 years ago, complete with Aero icons and skeuomorphic common controls.

Priorities? Sure, just not engineering excellence, attention to detail, or consistency in what the user sees.

Do I think this approach is used for Azure and Exchange security?

Now the key questions, “What engineering approach will Microsoft use as it applies smart large language models to Web search?”

Stephen E Arnold, January 5, 2022

Another Lilting French Cash Register Chime

January 2, 2023

An outfit call SC Magazine reported that the French cash registers — you know the quaint one with brass letters and the cheery red enamel — has chimed again. “Microsoft Fined $64 Million by France over Cookies Used in Bing Searches” reports:

France’s privacy watchdog fined Microsoft €60 million ($64 million) for not offering clear enough instruction for users to reject cookies used for online ads, as part of the move to enforce Europe’s tightening data protection law.

The write up noted:

Microsoft has been ordered to solve the issue within three months by implementing a simplified cookie refusal mechanism, or it could face additional fines of €60,000 a day…

It seems that some US companies do not take those French and EU regulations seriously. My suggestion to the Softies: France in not the US. Get on a couple of special lists and you may find some quality time in a glass room at CDG next time you visit. The good news is that US embassy personnel can visit you without too much red tape bedecking those gray suits.

Stephen E Arnold, January 2, 2023

In France, Tipping in Restaurants, Non. Showing Appreciation to the Government, Oui

December 23, 2022

Ah, France. Land of 200 cheeses, medallion bedecked chickens, and fat American high-tech creatures. Go to a French restaurant and order (in French certaInment) a cooked bird. Chow down. Do not tip the waiter. Say “merci” and smile. But if you a very fat, super large, very unpredictable American technology company tipping is mandatory.

Don’t believe me?

Read “Microsoft Hit with €60 Million Fine by France for Not Offering Cookie Opt Out on Bing.” Mais oui. The write up reports:

In addition, CNIL will fine Microsoft €60,000 per day within three months if it doesn’t ask users for consent to use an ad fraud detection cookie.

Will Microsoft’s paying up make the governmental doubt about Microsoft become like the mist in Verdon Gorge?

Ho ho ho.

In order to do business in France, American outfits have to go through a number of hoops. Some are easy; others require some bureaucratic finesse. One example is for an American company to sell something to the French government. There are hoops for American cheese. I have been informed that canned American cheese propelled by carbon dioxide is a hoot at some French parties. Mon dieu! Aerosol fromage. Interesting.

With the EU chasing some firms who say one thing and do another, fining some big tech companies is a way to get an allowance from mom. Amazon appears to decided to just pay up.

Microsoft may enter the fascinating French legal system to explain that its tracking devices are different. Oh, well. Some French judicial officials can use a mobile phone. But the cookie thing? Maybe not so much.

What’s the sound I hear? It is ka-ching.

Finding reasons to take legal action against US big tech companies is easy. The regret, as I understand it, is that it take a long time to get the money from the Americans.

What’s the outlook for 2023? That’s a softball question. The answer is more lawyers pecking on the confused Americans. The Monaco Grand Prix is in France right?

Stephen E Arnold, December 23, 2022

Microsoft Software Quality: Word Might Stop Working. No Big Deal

December 20, 2022

I read a short item which underscores my doubts about Microsoft’s quality methods. l have questions about security issues in Microsoft’s enterprise and cloud products and services. But those are mostly “new” and the Big Hope for future revenues. Perhaps games will arrive to make the Softies buy Teslas and beef up their retirement accounts, just not yet.

Microsoft Confirms Taskbar Bugs, Broken File Explorer, and App Issues in Windows 10” reports:

If you use Windows 10, you might experience the following symptoms:

  • ?The Weather or News and Interests widget or icons flickers on the Windows taskbar
  • ?The Windows taskbar stops responding
  • ?Windows Explorer stops responding
  • ?Applications including Microsoft Word or Excel might stop responding if they are open when the issue occurs

The weather and news are no big loss in my opinion. Microsoft believes that Windows 10 users want weather and news despite the mobile phone revolution. (Remember Microsoft and its play to create a mobile phone? Yeah, that was spun as fail early and fail fast. I think of that initiative as a basic fail, not a fast or early fail. Plain old fail.)

The Taskbar and file manager are slightly more interesting. A number of routine functions go south for some lucky Windows 10 users.

But the zinger fail is that Microsoft Word or Excel die. Now that’s just what’s needed to make the day of a person who is working on a report at a so-so consulting firm like one of the blue-chip outfits in Manhattan, a newbie at a big law firm with former government officials waiting for the worker bees to deliver a document for the bushy eyebrow set to review, or a Wall Street type modifying a model to make his, her, thems partners lots of money.

These happy users are supposed to be able to handle stress and pressure.

I wonder if Microsoft executives have been in a consulting firm, law firm, or financial services company when a must have app stops responding. Probably not because these wizards are working on improving Microsoft’s quality control processes. Could Redmond’s approach to quality be blamed on an intern, a contractor, or a part time worker? My hunch is that getting blamed is not a component of the top dogs’ job description.

Stephen E Arnold, December 20, 2022

« Previous PageNext Page »

  • Archives

  • Recent Posts

  • Meta