Chevron has hired Ellis to develop AI for seismic interpretation and earth modeling. Supposedly, it can do in a week what would take a team of earth scientists the better part of a year.
18 replies (most recent on top)
"AI", this years fad...
Better to replace management with AI: Much less complicated that petro-tech work!
3D geo rooms were the rage when I started work here. Fast forward to the best projects the company has and they are no where to be seen. But don’t worry someone probable got a Fellows out of it.
With 10 years to go, I’m not impressed or concerned. It’s great for writing my year-end reviews in workday, though.
Honestly, we have more to gain from pursuing automation of redundant work processes than expecting AI to produce tangible results at the push of a button. Keep dreaming. I remember when 3D seismic was going to answer all our geologic questions and get down to sub-inch resolution, uh, nope. Keep those geo-fantasies coming though.
Ok then, if after the first week this AI thing doesn’t prove itself to accomplish the annual workload of the entire team of earth scientists, get rid of it on Day 8, along with all the people who vouched it would succeed.
I wonder how much more money these goofballs are going to sink in so called digital? I personally know of at least 4 projects (>100MM spent) which didn't produce any tangible return that couldn't have been replicated by an excel spreadsheet. These are beautiful visualization fluff that no one knows what to do with or act upon (except for just saying "wow").
Back when we sent off some of our "best and brightest" to get mini-MIT degrees (rather than actually getting some work experience) and formed a study team of our "best and brightest" (but woefully inexperienced) earth scientists to study the possibility of AI "enhancing" (euphemism for replacing) seismic interpreters, this was always the stated goal. Boomers and experienced staff who questioned this were run off in 2020. Four years later, it's painfully obvious that our self-proclaimed "best and brightest" couldn't find oil at a refinery. Since the interpreter competency bar has now been set so low, it is indeed possible that AI could out-perform our current staff.
Has to be better then our current staff.
Ideally, AI will provide non-biased unemotional analysis. That would be a marked improvement. Secondly, no mind altering substances will be needed.
Lmao do you really understand what AI can do and/or cannot do in earth science? To actually think AI can magically deal with field geophysical data is just pure wishful thinking.
Could AI really do much worse than our current "success" rate?
Yes, dipsh-t management will be energized to replace experts with AI, the good management will realize the experts are enabled by AI to avoid repetitive tasks.
I think management has been heard to say that AI couldn't be any worse than the results we have seen from our expensive staff the last couple decades.
Perhaps AI will drill those new dry holes too. That said, autopicking was supposed to revolutionize mapping seismic volumes. For really nice clean high-fold seismic data with minimal noise and excellent frequency content, autopicking helps. Does anyone else remember that Aussie training volume we used back in the EPOS days? Haha, I mapped a really sweet time horizon that HP & KE would have drooled over with just a couple seeds and then never saw good data like that again. What I don't get is why the digital transformation didn't deliver on its promises. How many tech talks did we sit through where the speakers delivered promises that were totally pie-in-the-sky for any petrotech with a clue. Good luck!
I am sure MW and RM must be salivating at the prospect of laying off the last of the earth scientists. They have no interest in finding new reserves.
Management has long held the view that earth science, unlike engineering, is “expensive” relative to the value generated. This shortcut will have great appeal.
Maybe AI t could also generate comments for this board.
That sounds like a great sales pitch.