It feels like Nike is pushing everyone to use AI so they can figure out how to automate as many roles as possible. Eliminate those roles and then those at the new India campus will be doing a majority of the workload.
33 replies (most recent on top)
nothing gets past this one… must be from corp strat
There are some absolutely great engineers at Nike, but there are so many more that don't have the technical depth, personality, or the wherewithal to improve.
If anything, tech needs to cut another big chunk of people, and keep the ones that do right. We'd be faster and better aligned with less incapable hands clogging up the gears.
Literally going the way of Allbirds
my team does not even review the ai slop we push into prod
@mj What is that supposed to mean? As a publically traded company, decision makers have certain obligations, and that includes avoiding conflicts of interest. There is a reason the RAT was evicted as was being investigated for contracts that violated this.
@hz Shows how much you understand conflict of interest.
@dk this is a Nike wide problem ...very true for PHK as well. At any point in time, you can guarantee that there will be at least 3 teams working on the same problem...completely independently.
@hz You can say that about all the major tech solutions. Databricks, Snowflake, he-l, Service Now...
I wonder how many people signing the contracts for those AI tools properly disclosed their conflict of interest.
@gx The human glue is clogging up all the processes
Effective use of AI is built on clean, usable data and systemized institutional knowledge. There is so much human glue at Nike, I doubt AI can replace entire work flows, maybe it can speed up tasks in silos. Unless AI starts to read human minds. Then we're fu---d as a race anyway so Nike won't matter.
AI is nowhere near ability to automate a job. Now, it can definitely make someone more productive, but the fear of loosing your job to AI right or the near future now is a dog whistle.
@da didn’t Nike have a multi year effort called “manufacturing revolution” that was supposed to take care of this? Flexico anyone? What happened? 😂
Nike and AI? - teams are creating the same solutions in ITC that have already been created by Nike engineers in Phk with no leadership oversight or organization. Embarrassing the way AB is hyping ITC. They can’t even get the swoosh correct on AI presentations.
This will be deleted soon…..
When will they create fully automated shoemaking robot or machine?
@cr Huh? What are you guys talking about?
"While the company spent years in the red on raw infrastructure costs, recent financial disclosures from late 2026 show that Anthropic now generates $1.70 in revenue or every dollar spent on computing power (inference/serving costs), meaning they are making a gross profit on token delivery."
Yes, bq is full of sh-t with the $600k figure, but your guy here is literally inferring a gross profit per token just by looking at revenue-per-token and subtracting "compute expense per token", which currently only includes inference costs and hasn't been transparent on whether it accounts for land usage, electricity, maintenance over time, etc.
That is a pretty big distinction when talking about corporate financials that are primarily being disclosed to an audience of private investors.
@cq the "colocation model" is only going to make that harder, too.
they could be hiring 1 mid tier retail engineer in a nearby time zone, LCOL US area for every 3 vibe coding ITC guys they choose to hire instead, but instead they just 180'd on that opportunity and now to fit the locationally high PHK eng salary you're basically expected to larp as a temporarily embarrassed FAANG guy.
@cf I could ask the same. Is there any mention of net revenue or profit, only gross profit. No comparisons were made. Like, you go out of your way to be wrong twice while insulting the person you are failing to correct, it's hilarious.
The biggest problem is the folks on this forum think that Nike engineers are best in class. A lot of the engineers are below average from retail standards. So getting better talent is easier.
@cn Meant to say bq and not C1. These aliases are a bi--h to keep track of.
C1 is wrong, but it is true that big firms are getting sweetheart deals that are as good as 1/10th of what a normal user pays.
If the pursestrings from Anthropic start tightening, as they will, because they're still not at their expected cash flow targets, you're probably going to see spend per engineer double or triple, and that means hiring an army of cheap engineers to pump bloated codebases and paragraphs long copy pasted tickets full of user requirements they don't actually understand is going to cost more than just hiring one or two guys who can read and optimize his input token burn.
So Nike needs to get their head out of their as--s and not pretend that a bunch of cheap labor vibe-coding is going to give us huge cost savings.
There are already cooler heads trying to steer us out of that trap, but I don't have massive hopes that they'll prevail until I start seeing real change.
@c4 You seem confused, nobody is claiming that except you.
The only thing c1 is saying is that the variable costs are not the issue and they aren't losing money. Token limiting is more about shared capacity than cost control for per token spend. The point is bq's point is completely fabricated.
you might want to double check your accounting definitions if you think net revenue is profit
@c1 Anthropic didn't disclose their full expense profile in that statement, and real examples (like Accenture instituting token spend limits) suggest it's not the whole picture.
@bq Old information
While the company spent years in the red on raw infrastructure costs, recent financial disclosures from late 2026 show that Anthropic now generates $1.70 in revenue for every dollar spent on computing power (inference/serving costs), meaning they are making a gross profit on token delivery.
That requires leadership to listen instead of gutting orgs to replace them with $20k/yr "engineers" who can only subsist by cutting and pasting tickets into Claude lol
Possibly. Then it's up to those involved in applying to be vocal about what can't or shouldn't be automated.
A $6,000 spend on Claude is a $300,000 monthly cost to Anthropic. We're fu---d when they come knocking
@OP I mean, yeah? Agile reduces technological innovation to tickets. No thinking allowed. Then if it’s just tickets and no thinking, why can’t an AI prompt do it? Then if an AI prompt could do it, why not outsource the prompting?
I’ve already seen itc folks use Claude to do simple tickets they barely understand, and then use Claude to generate the update for the tickets, and even use slack to respond to emails and slack.
Then when nobody understands anything and sh-t comes crashing down, everyone responsible would have collected their millions and been off to the next one.
Directive is to deliver with “4x speed” with AI. My team has been pushing slop without even reviewing it anymore.
AI enables an organization to move more swiftly, not differently.
AI at Nike will improve the speed we lose market share. It will create an even bigger divide between senior positions and the unlucky ones that remain.
Think about the language used around the topic of ITC. Senior management looks at ITC like an appliance.
AI hastens the arrival to a destination predetermined by the prevailing trajectory.
It will enable increased speed to market of product that is now losing soul, identity, and relevance at an increasing rate.
If you thought senior managements over reliance on fresh consultants to determine incredibly important and irreversible business decisions was bad… imagine now those same senior management positions prompting things like “how make vested options more good, while less people and bad product? Also plz make Olympics plan”
Yes, you are absolutely correct. Nike is not the only company doing this.