Tavi
Yes, I mean why use things you expect to go wrong, often?
164 publicly visible posts • joined 17 Jan 2025
Having worked on JDE and unravelling a system , I would say Oracle is a bit better than SAP , and this gives us time to understand a system when we are also merging all police forces, Oracle were helpful with unravelling JDE , and gave us , well me insight to how the ERP works, so I think it's a positive in the world of pretend SaaS and AI vibe coding, people will learn something as they have to.
AI has allowed Ilegbodu to code in languages he doesn't yet know, such as Python, Bash, and Groovy.
But this context switching can get wearisome, he admitted. "At the end of the day, I'm actually kind of tired, because effectively, I spent the whole day talking to something."
So may as well be a chef then , with no understanding of heat, pectin etc ...
A familiar pivot: when the process falters, attention turns to something sweeter.
Shifting from ERP issues JDE or SAP ,to AI masks unresolved core instability.
I have seen this as poor data, fragile integrations, and inconsistent processes and they wonder why it undermine outcomes.
So AI with no technical knowledge how, essentially surface appeal over substance lets go with SaaS, so compounding risk.
Better to stabilise the base before refining what flows through it, it's not that hard , but you know this now.
Yep , and it's the locking that made us look first at azure services and functions, then the weather of knowledge on these and towards containerisation and maybe building an app that takes the 10% we use of everyone's SaaS , and making what we need , with actual developers, analysts and architects , savings are huge.
In the world of SF, the fluffy things that always pre-empt a sale, and the fluffy bears to take away and the multiple enterprises that use the software as they have no dev team, just over inflated C - Suite and the twat that got sold the salesforce cocaine.
In reality what this means is Salesforce introducing Agent Work Units (AWUs) suggests a move toward usage-based billing tied to AI output, or anything the want to say is AI <-- Open your Fkn Eyes!
So now your going to be billed on:
Harder cost predictability vs traditional seat licences.
AI workloads scale non-linearly ,one automation loop can multiply token usage (Moltbot style)
Finance teams lose fixed budgeting; spend becomes +/- Fuck Knows
And this means your IT teams are
Pay for activity, not outcomes.
Over-automation can create hidden operational complexity, like a 10'000 javascript page
IT Governance teams must monitor AI behaviour constantly, talk me through your last positon , oh, never mind!
In the land of the Fluffy Bear this most certainly means
Layoffs
Target "ServiceNow"
Heavy AI Positioning
Then automation, tokens, AWUs, Slack integration, and ITSM all converge (Esp with ServiceNowt)
Workflows become tightly coupled to Salesforce AI.
Migration costs increase sharply.
Customers lose negotiating leverage at renewal.
Eggs in one Benioff
So my gut is saying they have just lost the SaasPocalypse, in so far as we are developing our own tooling now its got a lot cheaper, and moving away from vendor lock-in , as you should be, expect the stock to take a hit soon.
We get AI to write us the parts of a SaaS we need for our company , then use a strangler fig pattern to get rid of the expensive SaaS employ a few more Devs and infra / data peeps with the cash and carry on doing this , use LLM as a corporate grass (as it's trained on SaaS)..
I think we are in business Randolph.
I have noticed the staff being tuped out, or in to Capgemini for years from HMRC , so don't expect any groundbreaking stuff, just keeping the sausage factory going and introducing complexity like kubrenetes and docker to modified cots systems , not actually replacing anything, I see pega is still there and some outdated content management systems , yes id enshittify these pronto to ensure tenure.