While his position may be a bit extreme, none of the panelists at the EAGE Annual Conference and Exhibition disagreed openly with Kristian Johansen (TGS CEO) who opined that, ‘The energy transition has failed completely – it is going to take much longer’. The buzz from the EAGE held last June in Toulouse France was a combination of schadenfreude at the failing transition, concern over the $70/barrel oil price and optimism that this would soon rise as demand increased. We will see. Meanwhile as the world economy teeters on the brink of sanctions what can we do here at Oil IT Journal but … carry on regardless. So here goes for a quick summary of our reporting in this, our 283rd edition.
AI continues to excite industry. TotalEnergies reports breathlessly on the ‘20,000 AI models’ currently in production. The OMG (now swallowed-up by the EDM Association) has rolled out a strategic playbook and AI Agent Capabilities Periodic Table. But not all agree – Gartner, Salesforce and especially Emily Bender are more nuanced on the AI ‘successes’. A PIDX webinar covering AI and the supply chain illustrated the overlap (in both reality and from the marketing department) of AI as per machine learning and ‘AI’ as per robotic process automation. If the latter is working why fix it? The answer probably is if you are selling RPA into a world addicted to machine learning models you may need to revisit your sales pitch. Conversely, pity the folks who have had ML embedded in their software tools for years who also need to tread carefully so as not to insult the buyers.
We have over the years reported on quantum computing. I would like to add ‘in oil and gas’ although to date, there is not much to report. Not least because a workable quantum computer is still on the drawing board. So it behooves us to report on a QC/algorithmic breakthrough from TotalEnergies which is set to change how energy companies optimize their portfolios. Well, that’s the idea. I always thought that portfolios were optimized by the CEO smooching with the local politicos but what do I know…
A report from the IOGP investigates the use of eye tracking technology to study a driller’s interaction with computer displays and the environment. Interestingly, drillers appear to be more interested in the numerical values displayed on screen than the fancy graphics. Which reminds me of Patrick Leach’s great book about oil and gas economics ‘Just give me the number’ that we reviewed many years ago.
In its Q2 quarterly results SLB enumerates a number of digital successes around AI, data and various applications. One might have expected some mention of OSDU, the open subsurface data universe, that SLB has sworn allegiance to, at least in private. But no, SLB, like Halliburton, studiously avoids any mention of the ‘collaboration-enabling’ OSDU in its official releases.
SLB has also teamed with France’s AI champion Mistral.ai as indeed has TotalEnergies. These days software and services providers (and indeed oil companies) tend to be what the Scots call ‘mealy mouthed’ when it comes to fossil energy. Instead of oil and gas, releases speak of ‘energy’ and the transition. Researchers work on designing batteries and optimizing wind turbine placement rather than oil and gas wells. We had an amusing conversation with Mistral’s Le Chat which kind of let the cat (ha ha) out of the bag before getting suspicious and displaying an suspicious degree of ‘woke’.
We were lucky enough to get a review copy of Esri’s booklet on GeoAI. As I said above, it’s tricky for companies that already do AI to market their wares as something new. Esri has had various bits and pieces of AI under the hood for years. So a rehash of marketing material is kind of natural. But what I would like to see debated is why the geospatial coordinates should be handled any differently from all the other inputs to an AI model. Why not just chuck everything and see what comes out?
We have a really big section on cyber security in this issue with plenty of warnings and advice for the oil and gas vertical, from insider risks, to the dangers of cloud-attached energy assets. Protection is available in the form of a plethora of associations and standards. By the time a CIO has aligned with all the standards a thousand script kiddies and other bad actors have downloaded as many rootkits, malwares and what have you. This is looking like a losing battle!
In our Standards stuff section we report on the usual standards bulimics the IOGP, the OGC and others. One standard that was new to us is the Model Context Protocol, a means of tying various LLMs and other AI models together in an ‘agentic’ framework. Interoperability at last? Maybe too soon to tell. A paper in Nature of all places investigates data lineage tracking by combining ‘semantic technologies’ with logical principles. Talk about the standards moral high ground! Maybe time for OSDU to retool? Another ‘we do AI already’ in our Software short takes rubric from Wolfram/Mathematica. Agentics, maps, geodetics and a vibe code generator for Wolfram, it’s all there! Finally in our exclusive interview with John Bottega, CEO of the EDM Association we get the inside track of the acquisition, or was it a merger, with the OMG. Enjoy. Well, ‘enjoy’ if you subscribe.
© Oil IT Journal - all rights reserved.