Dumb Smart Software? This Is News?
January 31, 2025
A blog post written by a real and still-alive dinobaby. If there is art, there is AI in my workflow.
The prescient “real” journalists at the Guardian have a new insight: When algorithms are involved, humans get the old shaftola. I assume that Weapons of Math Destruction was not on some folks’ reading list. (O’Neil, Cathy. Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy. New York: Crown, 2016). That book did a reasonably good job of explaining how smart software’s math can create some excitement for mere humans. Anecdotes about Amazon’s management of its team of hard-working delivery professionals shifting into survival tricks revealed by the wily Dane creating Survival Russia videos for YouTube.
(Yep, he took his kids to search for graves near a gulag.) “It’s a Nightmare: Couriers Mystified by the Algorithms That Control Their Jobs” explains that smart software raises some questions. The “real” journalist explains:
This week gig workers, trade unions and human rights groups launched a campaign for greater openness from Uber Eats, Just Eat and Deliveroo about the logic underpinning opaque algorithms that determine what work they do and what they are paid. The couriers wonder why someone who has only just logged on gets a gig while others waiting longer are overlooked. Why, when the restaurant is busy and crying out for couriers, does the app say there are none available?
Confusing? To some but to the senior managers of the organizations shifting to smart software, the cost savings are a big deal. Imagine. In Britain, a senior manager can spend a week or two in Nice, maybe Monaco? The write up reports:
The app companies say they do have rider support staffed by people and some information about the algorithms is available on their websites and when drivers are initially “onboarded”.
Of course the “app companies” say positive things. The issue is that management embraces smart software. A third-party firm is retained to advise the lawyers and accountants and possibly one presentable information technology person to a briefing. The options are considered and another third-party firm is retained to integrate the smart software. That third-party retains a probably unpresentable IT person who can lash up some smart software to the bailing-wire-and-spit enterprise software system. Bingo! The algorithms perform their magic. Oh, whom does one blame for a flawed solution? I don’t know. Just call in the lawyers.
The article explains the impact on a worker who delivers for people who cannot walk to a restaurant or the grocery:
“Every worker should understand the basis on which they are paid,” Farrar [a delivery professional] said. “But you’re being gamed into deciding whether to accept a job or not. Will I get a better offer? It’s like gambling and it’s very distressing and stressful for people. You are completely in a vacuum about how best to do the job and because people often don’t understand how decisions are being made about their work, it encourages conspiracies.”
To whom should Mr. Farrar and others shafted by math complain? Perhaps the Guardian newspaper, which is slightly less popular than TikTok or X.com, Facebook or Red Book, or BlueSky or YouTube. My suggestion would be for the Guardian to use these channels and beg for pounds or dollars like other valiant social media professionals. The person doing deliveries might want to explore working for Amazon deliveries and avail himself of Survival Russia videos when on his generous Amazon breaks. And what about the people who call a restaurant and specify at home delivery? I would recommend getting out of that comfy lounge chair and walking to the restaurant in person. While you wait for your lovingly-crafted meal at the Indian takeaway, you can read Weapons of Math Destruction.
Stephen E Arnold, January 31, 2025
Two Rules for Software. All Software If You Can Believe It
January 31, 2025
Did you know that there are two rules that dictate how all software is written? No, we didn’t either. FJ van Wingerde from the Ask The User blog states and explains what the rules are in his post: “The Two Rules Of Software Creation From Which Every Problem Derives.” After a bunch of jib jab about the failures of different codes, Wingerde states the questions:
“It’s the two rules that actually are behind every statement in the agile manifesto. The manifesto unfortunately doesn’t name them really; the people behind it were so steeped in the problems of software delivery—and what they thought would fix it—that they posited their statements without saying why each of these things are necessary to deliver good software. (Unfortunately, necessary but not enough for success, but that we found out in the next decades.) They are [1] Humans cannot accurately describe what they want out of a software system until it exists. and [2] Humans cannot accurately predict how long any software effort will take beyond four weeks. And after 2 weeks it is already dicey.”
The first rule is a true statement for all human activities, except the inability to accurately describe the problem. That may be true for software, however. Humans know they have a problem, but they don’t have a solution to fix. The smart humans figure out how to solve the problem and learn how to describe it with greater accuracy.
As for number two, is project management and weekly maintenance on software all a lucky guess then? Unless effort changes daily and that justifies paying software developers. Then again, someone needs to keep the systems running. Tech people are what keep businesses running, not to mention the entire world.
If software development only has these two rules, we now know why why developers cannot provide time estimates or provide assurances that their software works as leadership trained as accountants and lawyers expect. Rest easy. Software is hopefully good enough and advertising can cover the costs.
Whitney Grace, January 31, 2025
Happy New Year the Google Way
January 31, 2025
We don’t expect Alphabet Inc. to release anything but positive news these days. Business Standard reports another revealing headline, especially for Googlers in the story: "Google Layoffs: Sundar Pichai Announced 10% Job Cuts In Managerial Roles.” After a huge push in the wake of wokeness to hire under represented groups aka DEI hires, Google has slowly been getting rid of its deadweight employees. That is what Alphabet Inc. probably calls them.
DEI hires were the first to go, now in the last vestiges of Googles 2024 push for efficiency, 10% of its managerial positions are going bye-bye. Among those positions are directors and vice presidents. CEO Sundar Pichai says the push for downsizing also comes from bigger competition from AI companies, such as OpenAI. These companies are challenging Google’s dominance in the tech industry.
Pichai started the efficiency push in 2022 when people were starting to push back against the ineffectiveness of DEI hires, especially when their budgets were shrunk from inflation. In January 2023, 12,000 employees were laid off. Picker is also changing the meaning of “Googleyness”:
“At the same meeting, Pichai introduced a refined vision for ‘Googleyness’, a term that once broadly defined the traits of an ideal Google employee but had grown too ambiguous. Pichai reimagined it with a sharper focus on mission-driven work, innovation, and teamwork. He emphasized the importance of creating helpful products, taking bold risks, fostering a scrappy attitude, and collaborating effectively. “Updating modern Google,” as Pichai described it, is now central to the company’s ethos.”
The new spin on being Googley. Enervating. A month into the bright new year, let me ask a non Googley question: “How are those job searches, bills, and self esteem coming along?
Whitney Grace, January 31, 2025
AI Innovation: Writing Checks Is the Google Solution
January 30, 2025
A blog post from an authentic dinobaby. He’s old; he’s in the sticks; and he is deeply skeptical.
Wow. First, Jeff Dean gets the lateral arabesque. Then the Google shifts its smart software to the “I am a star” outfit Deep Mind in the UK. Now, the cuddly Google has, according to Analytics India, pulled a fast one on the wizards laboring at spelling advertising another surprise. “Google Invests $1 Bn in Anthropic” reports:
This new investment is separate from the company’s earlier reported funding round of nearly $2 billion earlier this month, led by Lightspeed Venture Partners, to bump the company’s valuation to about $60 billion. In 2023, Google had invested $300 million in Anthropic, acquiring a 10% stake in the company. In November last, Amazon led Anthropic’s $4 billion fundraising effort, raising its overall funding to $8 billion for the company.
I thought Google was quantumly supreme. I thought Google reinvented protein stuff. I thought Google could do podcasts and fix up a person’s Gmail. I obviously was wildly off the mark. Perhaps Google’s “leadership” has taken time from writing scripts for the Sundar & Prabhakar Comedy Tour and had an epiphany. Did the sketch go like this:
Prabhakar: Did you see the slide deck for my last talk about artificial intelligence?
Sundar: Yes, I thought it was so so. Your final slide was a hoot. Did you think it up?
Prabhakar: No, I think little. I asked Anthropic Claude for a snappy joke. It worked.
Sundar: Did Jeff Dean help? Did Dennis Hassabis contribute?
Prabhakar: No, just Claude Sonnet. He likes me, Sundar.
Sundar: The secret of life is honesty, fair dealing, and Code Yellow!
Prabhakar: I think Google intelligence may be a contradiction in terms. May I requisition another billion for Anthropic?
Sundar: Yes, we need to care about posterity. Otherwise, our posterity will be defined by a YouTube ad.
Prabhakar: We don’t want to take it in the posterity, do we?
Sundar: Well….
Anthropic allegedly will release a “virtual collaborator.” Google wants that, right Jeff and Dennis? Are there anti-trust concerns? Are there potential conflicts of interest? Are there fears about revenues?
Of course not.
Will someone turn off those darned flashing red and yellow lights! Innovation is tough with the sirens, the lights, the quantumly supremeness of Googleness.
Stephen E Arnold, January 30, 2025
Who Knew? A Perfect Bribery Vehicle, According to Ethereum Creator
January 30, 2025
A blog post from an authentic dinobaby. He’s old; he’s in the sticks; and he is deeply skeptical.
I read “Ethereum Creator Vitalik Buterin: Politician Issued Coins Perfect Bribery Vehicle.” Isn’t Mr. Buterin a Russian Canadian? People with these cultural influences can spot a plastic moose quickly in experience.
The write up reports:
Ethereum founder Vitalik Buterin has criticized cryptocurrencies issued by politicians as “a perfect bribery vehicle.” “If a politician issues a coin, you do not even need to send them any coins to give them money,” Buterin explained in a tweet. “Instead, you just buy and hold the coin, and this increases the value of their holdings passively.” He added that one of the reasons these “politician coins” are potentially excellent tools for bribery is the element of “deniability.”
Mr. Buterin is quoted in the write up as saying:
“I recommend politicians do not go down this path.”
Who knew that a plastic moose would become animated and frighten the insightful Russian Canadian? What sound does a plastic moose make? Hee haw hee haw.
Nope, that’s a jackass. Easy mistake.
Stephen E Arnold, January 30, 2025
Ah, the Warmth of the Old, Friendly Internet. For Real?
January 30, 2025
I never thought I’d be looking back at the Internet of yesteryear nostalgically. I hated the sound of dial-up and the instant messaging sounds were annoying. Also AOL had the knack of clogging up machines with browsing history making everything slow. Did I mention YouTube wasn’t around? There are somethings that were better in the past, including parts of the Internet, but not all of it.
We also like to think that the Internet was “safer” and didn’t have predatory content. Wrong! Since the Internet’s inception, parents were worried about their children being the victims of online predators. Back then it was easier to remain anonymous, however. El País agrees that the Internet was just as bad as it is today: “‘The internet Hasn’t Made Us Bad, We Were Already Like That’: The Mistake Of Yearning For The ‘Friendly’ Online World Of 20 Years Ago."
It’s strange to see artists using Y2K era technology as art pieces and throwbacks. It’s a big eye-opener to aging Millennials, but it also places these items on par with the nostalgia of all past eras. All generations love the stuff from their youth and proclaim it to be superior. As the current youth culture and even those middle-aged are obsessed with retro gear, a new slang term has arisen: “cozy tech.”
“‘Cozy tech’ is the label that groups together content about users sipping from a steaming cup, browsing leisurely or playing nice, simple video games on devices with smooth, ergonomic designs. It’s a more powerful image than it seems because it conveys something we lost at some point in the last decade: a sense of control; the idea that it is possible to enjoy technology in peace again.”
They’re conflating the idea with reading a good book or listening to music on a record player. These “cozy tech” people are forgetting about the dangers of chatrooms or posting too much information on the Internet. Dare we bring up Omegle without drifting down channels of pornography?
Check out this statement:
“Mayte Gómez concludes: “We must stop this reactionary thinking and this fear of technology that arises from the idea that the internet has made us bad. That is not true: we were already like that. If the internet is unfriendly it is because we are becoming less so. We cannot perpetuate the idea that machines are entities with a will of their own; we must take responsibility for what happens on the internet.”
Sorry, Mayte, I disagree. Humans have always been unfriendly. We now have a better record of it.
Whitney Grace, January 30, 2025
Microsoft and Security: What the Style Guide Reveals
January 29, 2025
A blog post written by a real and still-alive dinobaby. If there is art, there is AI in my workflow.
If you have not seen the “new” Microsoft style guide, you will want to take a quick look. If you absorb the document, you might qualify to write words for Microsoft. I know that PR and marketing are important. We have had some fun trying to get Visio to print after the latest unwanted update. Some security issues exist for a number of Microsoft products and services. Do you want “salt” on your solar windburn?
With security in mind, I wanted to see what the style guide offers the Microsoftie trying to learn. Using the provided search system, I saw 29 entries about security. These came from “documentation.” Yep, zero references to security, how to handle it, what to say, the method of presenting security information, nada. A Microsoftie or a curious dinobaby like me would not see the word security in the style guide’s information for:
- Credentials
- Q&A
- Reference
- Shows
- Training
Was I surprised? No, a style guide is not focused on security. But I think some discussion of the notion of security and how to respond when an all-to-frequent breach is discovered would be useful. I make this remark because the top dogs of Microsoft said security was Job One … at least until AI, Copilot, and trying to recoup costs became Job One with a gold star.
Does anyone care? Not too much.
Stephen E Arnold, January 29, 2025
How Does Smart Software Interpret a School Test
January 29, 2025
A blog post from an authentic dinobaby. He’s old; he’s in the sticks; and he is deeply skeptical.
I spotted an article titled “‘Is This Question Easy or Difficult to You?’: This LSAT Reading Comprehension Question Is Breaking Brains.” Click bait? Absolutely.
Here’s the text to figure out:
Physical education should teach people to pursue healthy, active lifestyles as they grow older. But the focus on competitive sports in most schools causes most of the less competitive students to turn away from sports. Having learned to think of themselves as unathletic, they do not exercise enough to stay healthy.
Imagine you are sitting in a hot, crowded examination room. No one wants to be there. You have to choose one of the following solutions.
(a) Physical education should include noncompetitive activities.
[b] Competition causes most students to turn away from sports.
[c] People who are talented at competitive physical endeavors exercise regularly.
[d] The mental aspects of exercise are as important as the physical ones.
[e] Children should be taught the dangers of a sedentary lifestyle.
Okay, what did you select?
Well, the “correct” answer is [a], Physical education should include noncompetitive activities.
Now how did some of the LLMs or smart software do?
ChatGPT o1 settled on [a].
Claude Sonnet 3.5 spit out a page of text but did conclude that the correct answer as [a].
Gemini 1.5 Pro concluded that [a] was correct.
Llama 3.2 90B output two sentences and the correct answer [a]
Will students use large language models for school work, tests, and real life?
Yep. Will students question or doubt the outputs? Nope.
Are the LLMs “good enough”?
Yep.
Stephen E Arnold, January 29, 2025
The Joust of the Month: Microsoft Versus Salesforce
January 29, 2025
These folks don’t seem to see eye to eye: Windows Central tells us, “Microsoft Claps Back at Salesforce—Claims ‘100,000 Organizations’ Had Used Copilot Studio to Create AI Agents by October 2024.” Microsoft’s assertion is in response to jabs from Salesforce CEO Marc Benioff, who declares, “Microsoft has disappointed everybody with how they’ve approached this AI world.” To support this allegation, Benioff points to lines from a recent MarketWatch post. A post which, coincidentally, also lauds his company’s success with AI agents. The smug CEO also insists he is receiving complaints about his giant competitor’s AI tools. Writer Kevin Okemwa elaborates:
“Benioff has shared interesting consumer feedback about Copilot’s user experience, claiming customers aren’t finding themselves transformed while leveraging the tool’s capabilities. He added that customers barely use the tool, ‘and that’s when they don’t have a ChatGPT license or something like that in front of them.’ Last year, Salesforce’s CEO claimed Microsoft’s AI efforts are a ‘tremendous disservice’ to the industry while referring to Copilot as the new Microsoft Clippy because it reportedly doesn’t work or deliver value. As the AI agent race becomes more fierce, Microsoft has seemingly positioned itself in a unique position to compete on a level playing field with key players like Salesforce Agentforce, especially after launching autonomous agents and integrating them into Copilot Studio. Microsoft claims over 100,000 organizations had used Copilot Studio to create agents by October 2024. However, Benioff claimed Microsoft’s Copilot agents illustrated panic mode, majorly due to the stiff competition in the category.”
One notable example, writes Okemwa, is Zuckerberg’s vision of replacing Meta’s software engineers with AI agents. Oh, goodie. This anti-human stance may have inspired Benioff, who is second-guessing plans to hire live software engineers in 2025. At least Microsoft still appears to be interested in hiring people. For now. Will that antiquated attitude hold the firm back, supporting Benioff’s accusations?
Mount your steeds. Fight!
Cynthia Murrell, January 29, 2025
What Do DeepSeek, a Genius Girl, and Temu Have in Common? Quite a Lot
January 28, 2025
A write up from a still-living dinobaby.
The Techmeme for January 28, 2024, was mostly Deepseek territory. The China-linked AI model has roiled the murky waters of the US smart software fishing hole. A big, juicy AI creature has been pulled from the lake, and it is drawing a crowd. Here’s a small portion of the datasphere thrashing on January 28, 2025 at 0700 am US Eastern time:
I have worked through a number of articles about this open source software. I noted its back story about a venture firm’s skunk works tackling AI. Armed with relatively primitive tools due to the US restriction of certain computer components, the small team figured out how to deliver results comparable to the benchmarks published about US smart software systems.
Genius girl uses basic and cheap tools to repair an old generator. Americans buy a new generator from Harbor Freight. Genius girl repairs old generator proving the benefits of a better way or a shining path. Image from the YouTube outfit which does work the American way.
The story is torn from the same playbook which produces YouTube “real life” stories like “The genius girl helps the boss to repair the diesel generator, full of power!” You can view the one-hour propaganda film at this link. Here’s a short synopsis, and I want you to note the theme of the presentation:
- Young-appearing female works outside
- She uses primitive tools
- She takes apart a complex machine
- She repairs it
- The machine is better than a new machine.
The videos are interesting. The message has not been deconstructed. My interpretation is:
- Hard working female tackles tough problem
- Using ingenuity and hard work she cracks the code
- The machine works
- Why buy a new one? Use what you have and overcome obstacles.
This is not the “Go west, young man” or private equity approach to cracking an important problem. It is political and cultural with a dash of Hoisin technical sauce. The video presents a message like that of “plum blossom boxing.” It looks interesting but packs a wallop.
Here’s a point that has not been getting much attention; specifically, the AI probe is designed to direct a flow of energy at the most delicate and vulnerable part of the US artificial intelligence “next big thing” pumped up technology “bro.”
What is that? The answer is cost. The method has been refined by Shein and Temu by poking at Amazon. Here’s how the “genius girl” uses ingenuity.
- Technical papers are published
- Open source software released
- Basic information about using what’s available released
- Cost information is released.
The result is that a Chinese AI app surges to the top of downloads on US mobile stores. This is a first. Not even the TikTok service achieved this standing so quickly. The US speculators dump AI stocks. Techmeme becomes the news service for Chinese innovation.
I see this as an effective tactic for demonstrating the value of the “genius girl” approach to solving problems. And where did Chinese government leadership watch the AI balloon lose some internal pressure. How about Colombia, a three-hour plane flight from the capital of Central and South America. (That’s Miami in the event my reference was too oblique.)
In business, cheaper and good enough are very potent advantages. The Deepseek AI play is indeed about a new twist to today’s best method of having software perform in a way that most call “smart.” But the Deepseek play is another “genius girl” play from the Middle Kingdom.
How can the US replicate the “genius girl” or the small venture firm which came up with a better idea? That’s going to be touch. While the genius girl was repairing the generator, the US AI sector was seeking more money to build giant data centers to hold thousands of exotic computing tools. Instead of repairing, the US smart software aficionados were planning on modular nuclear reactors to make the next-generation of smart software like the tail fins on a 1959 pink Cadillac.
Deepseek and the “genius girl” are not about technology. Deepseek is a manifestation of the Shein and Temu method: Fast cycle, cheap and good enough. The result is an arm flapping response from the American way of AI. Oh, does the genius girl phone home? Does she censor what she says and does?
Stephen E Arnold, January 28, 2025
-
- Subscribe to Beyond Search
Feature archive
News archive
-
