He began with some big claims about the impact of technology on productivity, siting impressive levels of GDP attributable to the tech sector and a claim that tech has driven multi-factor productivity (MFP) growth. The latter is an interesting claim since way back in 1987 Nobel laureate Robert Solow quipped 'You can see the computer age everywhere but in the productivity statistics.' This effect later termed the productivity paradox persisted through the 1970s and 80s, but partly resolved in the 90s. However, in this century the effect has been reidentified, this time as the productivity slowdown.
Has the ATA found real evidence. Unfortunately not. They have relied on a Department of Infrastructure paper that only identified the technology sector itself as having a higher MFP performance than the rest of the economy. That does not indicate that technology has driven MFP growth in the economy as a whole. This and other statistics cited certainly provide evidence for the claim that if the technology sector wasn't exhibiting such strong MFP growth then economy wide MFP growth would be lower - but not because MFP growh would be lower in those sectors.
As a champion of the NBN and better communications infrastructure in general, I accept that telecommunications (but not necessarily the broader tech sector) is an essential General Purpose Technology to enable growth. But it is only an enabler, not a driver.
The productivity slowdown has, in part, been blamed on the fact that GDP statistics do not capture the increased quality of services in many sectors. The ATA does not miss the opportunity to apply this argument to telecommunications. Just a small note - you cannot simultaneously rely on a statistic and then doubt its validity. For each reference to the 'quality' argument you could match it with the 'well being' argument, or the fact that GDP only measures aggregates and not distribution. What does it benefit the average Australian if the income of the top 1% grows by 40% but the income of the rest by zero?
The poor arguments take another leap when the subject moves to AI. Firstly we are being pitched the importance of AI being accessed through mobile devices, and then the benefits of sovereign capability in digital infrastructure. Sorry, but a whole lot of data centrs being owned by Australians to house services provided by global tech behemoths does not equate to sovereign capability. In all the discussion on data centres I haven't seen analysis that suggest any of these are for anything other than domestic use. Are consumers and businesses in South East Asia going to be accessing AI tools from Australian data centres? Is it a viable alternative for Australians to be accessing AI tools in data centres in South East Asia or China?
We are then treated to another version of 'the end is nigh' for telcos who are supposedly investing in projects with ROIs below cost of capital. I have never met a CEO or CFO who would allow that. The only value destroying investment I recall was Telstra's investment in Pay TV infrastructure - but that was only because the alternative resukted in greater value destruction.
Date growt may well be outstripping revenue growth, but the telco pricing in an oligopoly. Whether it is more Bertrand rather tha Cournot is a matter for debate, there is the possibility of the former because the nature of fibre optic transmission has meant that till now there have been few capacity constraints. However, the rate of capacity improvement on a single fibre is now slowing while the rate of traffic increase is staying the same. But this simply makes the case that the era of destructive competition - if that is what we had - is coming to an end.
The ATA repeated the moans about excessive regulation. They rattled off the 3,000 land access activity notices required by one telco to build a fibre route. Which of those land holders does the ATA suggest should NOT have had to receive a notice. Indeed the easiest way to reduce the cost of regulation is to stop duplicating infrastructure. This is the cost of infrastructure based competition which has been the touchstone of policy for so long. The NBN demonstrated that monopoly access infrastructure makes sense. It is time to reconsider whether monopoly transmission infrastructure wouldn't also make sense.
Q&A from the session was partly reported in Comms Day to which I don't subscribe, but some aspects were included in an email to TelSoc members. There were two interesting parts to this - one on sovereignty and one on 6G.
Coleman observed 'You no longer need to have fibre optic cables in the dirt here in Australia or mobile towers here in Australia to provide connectivity to Australians.' The same has been true ever since the first geostationary satellies were launched if we accept the dishes at each end can be consumer owned. The same was true of private microwave links. The trick is that the Telecommunications Act only deals with services to the public. To provide a service the LEO operators do need ground station infrastructure.
That an emergency call service obligation might accrue to a person who has no infrastructure in Australia is not new - satellite phones are not new. The trick is to make sure that the entity that is providing the service to the consumer is required to contractually procure the necessary services.
The 6G issue is far far more interesting. The progression through the Gs has been fascinating. First generation (analog) cellular services were bespoke in many countries and dated from the era of mostly single monopoly operators (except in the US). 2G (the first digital services) took an interesting bend - all of Europe cooperated in mandating GSM wile the US fragmented between TDMA and CDMA and a touch of GSM. Here Australia followed Europe and NZ followed the US.
The launch of 3G saw an accomodation between the GSM (or 3GPP) worldview and the CDMA worldview. Technically 3G used the air-interface standards developed for 3G from the CDMA world together with the network core standards from the GSM world. Spectrum was till mostly bands historically allocated.
4G also resulted from an accomodation - this time of the MIMO technology developed for WiMax with an evolution of the 3G core. This also coincided with the refarming of 'digital dividend' spectrum from television. 5G was the first standard that seemed to be developed from day one as a single 'collaborative' project. In reality the standard is a compliation of many bits of IP developed by different players in the eco system. It is also the first that has offered genuinely new capability at the edge so that services locally could operate even if connection to the core was lost. It also promised spectrum agility.
6G seems to have aspirations for so much more - but discussion of inherent AI capability doesn't make much sense. AI is just a technology that build gigantic adjacency matrices between objects (which can be works or images) and then uses those matrices to generate new material by using the adjacency matrices. Certainly the model can be subsetted down to smaller domains.
Despite the aspiration for more from 6G the Next Generation Mobile Networks Alliance (NGMN Alliance) has written in its position statement:
6G must not inherently trigger a hardware refresh of 5G RAN infrastructure. The decision to refresh 5G RAN hardware for operational reasons such as end-oflife, energy consumption or new capabilities must be an operator driven choice, independent of supporting 6G.
In short, elsewhere around the globe the network operators are pushing back on being bullied by the equipmnt industry into another full network refresh.
However, I do agree with the ATA that Australian Spectrum Management isn't up to speed for meeting the needs of future mobiles. However, I lay the blame for that on the policy approavh which has perpetuated the myth that spectrum auctions find the highest and best use for spectrum. The potential uses of spectrum are determined by equipment decisions, which in turn are driven by the ITU. The ATAT shouldnt be putting pressure on the ACMA to make a planing decision, they should be pressuring the Minister to change the Radiocommunications Act.
If only the ATA's positions weren't so blatanly industry positions being peddled as big national policy issues they could be worth lisening to. Charles Wilson was famously misquorted as saying 'What's good for General Motors is good for the country.' Unfortunately one feels that Australia's telcos believe the statement is true with their ow names inserted.
PS Coleman was reportedly 'supportive of NBN Co’s move from its Sky Muster geostationary satellites towards wholesale Amazon LEO services'. This is a pity - only because NBNCo has chosen the wrong LEO provider and that Australia could have gone much further earlier and become a co-owner of Starlink and acquired the rights for all of Australia and the Pacific. This was suggested in various RTIRC reports and submissions.
PPS As is so frequently the case, Australia has no direct presence in the NGMN Alliance - though Vodafone is a member. SK Telecom and China Telecom are both members.
PPPS Wilson's actual statement was made during confirmation hearings for appointment as the Secretary of Defense. When asked if he could make a decision as Secretary of Defense that would be adverse to the interests of General Motors, Wilson answered affirmatively. But he added that he could not conceive of such a situation "because for years I thought what was good for our country was good for General Motors, and vice versa."
*********************************
Life is what happens while you are busy making other plans JWL
*********************************
Life is what happens while you are busy making other plans JWL
No comments:
Post a Comment