Nine to Noon Notes 29 10 26
Here are my notes from today’s appearance on RNZ Nine to Noon show.
Listen online here: www.rnz.co.nz/national/…
Joint venture between One NZ and 2degrees
One NZ and 2degrees want to set up a joint venture that would combine their mobile radio networks while allowing them to continue competing as separate companies.
The Commerce Commission is investigating what this could mean for competition.
The joint venture, RanCo, would own and operate the companies’ active radio access network equipment. This includes the antennas and base stations.
One NZ and 2degrees would each buy network services from the new business but retain control of their own spectrum rights, core networks, fibre backhaul, satellite services and retail operations. They would continue competing on price, products, service and innovation. The proposal is subject to regulatory approval, with completion planned for the first half of 2027.
The key question is whether the companies can remain effective competitors when they no longer operate independent radio access networks. Not everyone thinks they can.
The companies argue that they will retain the assets and business decisions needed to differentiate themselves. Opponents question whether that separation will be sufficient, particularly when it comes to network investment, quality and the incentives to compete.
I’ve looked at the submissions to the Commission. Some warn that we risk ending up with the same problems as the supermarket sector: two dominant players, high prices and little room for new entrants.
Others fear the joint venture could be a first step towards a full merger between the two companies. Another submission argues that the deal would help 2degrees compete more effectively.
The arrangement also raises questions about how much each company will learn about the other’s network. Sharing infrastructure and its operation could give them access to commercially sensitive information about capacity, costs and investment plans. That could affect how effectively they compete, even if they remain separate businesses.
Reverse engineer anything
AI is making it much easier to take apart existing software, reproduce its features and modify devices that manufacturers have locked down or abandoned.
What was once a specialist activity requiring considerable technical knowledge is becoming a hobby. A project called Reverse Engineer Anything on GitHub, promises to let AI agents examine software without access to its source code, explain how features work and build copies.
People are using AI to strip advertising and unwanted software from smart TVs, simplify their menus and revive gadgets manufacturers have bricked or stopped supporting. One guitar-pedal maker claimed to have reproduced the effects algorithms of three competitors in 72 hours, potentially allowing him to sell copies of their products.
There are clear benefits. Consumers can regain control of the devices they have paid for, restore abandoned functionality and escape software clutter. But the same tools make it easier to copy other people’s work without permission, raising questions about intellectual property and the viability of businesses that depend on proprietary software.
There are also technical risks. AI-generated copies often contain bugs or reproduce only part of the original functionality. People using them may lack the expertise to identify and fix problems. Poorly written, insecure code could end up running on internet-connected devices, creating opportunities for hackers to introduce malware.
The biggest commercial threat may be to smaller software and hardware companies. Competitors could reproduce products that took months or years to develop, then undercut their creators. The result could be a flood of cheap, unreliable copies alongside genuinely useful modifications that give consumers more control over their devices.
AI not delivering promised payoffs for businesses
A prominent US investment analyst says artificial intelligence has yet to deliver measurable financial benefits for most companies outside the technology sector.
Torsten Slok, chief economist at Apollo Global Management, says revenue per employee and profit margins are rising among the seven largest US technology companies but remain flat across the other 493 companies in the S&P 500 index. He also sees no clear evidence of an AI-driven lift in productivity growth.
The argument echoes a debate dating back to the 1980s. Economist Robert Solow famously observed in 1987 that computers were everywhere except in the productivity statistics.
Despite decades of investment in information technology, productivity growth remained sluggish. The picture changed in the late 1990s, when computers, cheaper processing power and better business systems helped lift US productivity.
However, the expected economy-wide gains from subsequent waves of digital technology have proved less impressive. The OECD says digitalisation has improved productivity, but the gains have not been large enough to offset broader economic headwinds.
New Zealand faces its own productivity challenge. Almost all local businesses use digital tools, while MBIE’s 2026 research found that half believe they would benefit from using them more effectively. The question is whether AI will deliver measurable improvements in revenue, costs and output rather than simply add another item to technology budgets.
Slok argues that AI’s commercial value depends on whether ordinary businesses can turn increasingly cheap computing intelligence into higher revenue and profit. That requires judgement, organisational change and accountability, not just access to powerful models. The technology may be impressive, but the business case still needs to be proved.
Slok’s claim that AI has delivered “zero financial payoff” is based on the evidence of mainstream US companies, He isn’t saying that there has been. no businesses benefit.
New Zealand research commissioned by 2degrees has reported higher revenue among businesses adopting AI, although that association alone does not prove AI caused the difference.
Meta faces record fine for data privacy lies
A New Mexico judge has ordered Meta, the company behind Facebook, Instagram and WhatsApp, to pay US$25 billion after a jury found it misled users about how Facebook handled their personal data. The case followed the Cambridge Analytica scandal, in which a political consultancy obtained information from about 87 million Facebook profiles for political targeting. The judge found that misleading statements continued from 2010 to 2021 and involved senior executives, including Mark Zuckerberg.
The penalty was calculated at US$219 billion under state law before the judge reduced it to US$25 billion. Meta says it will appeal. The judgment follows a US$5 billion Federal Trade Commission settlement in 2019 and separate US litigation over child safety, including a New Mexico judgment of about US$942 million.
The case illustrates the growing legal risks facing technology companies whose business models depend on collecting and using personal data. Meta and other large technology companies have also become increasingly active in US politics, lobbying for less regulation and legal liability. Critics argue that their political spending is partly intended to limit the laws and enforcement measures that hold them accountable. The New Mexico judgment is not final until Meta pursues its appeal.
AI is not delivering promised payoffs for businesses
This is something I have been wondering about for some time.
A prominent US investment analyst writes:
Despite massive investments and widespread hype around artificial intelligence, the technology has delivered zero financial payoff to corporate America outside of tech.
We’ve been here before with earlier digital technologies.
In 1987 Nobel prize winning economist Robert Solow said:
We see computers everywhere except in the productivity statistics.
Computers had been commercially available since the 1950s, and personal computers became widespread in the 1980s. Nevertheless, the anticipated productivity gains were slow to appear.
Now there was some evidence of a pay off a decade later in the 1990s, but not before or after. It’s not entirely clear that was from IT, it could have been some other economic effect.