Some Family thing/some work thing/some pronoun thing/I paid for grok
Joined July 2018
- Tweets12.5K
- Following501
- Followers177
- Likes23.8K
Yup. I mentioned in another thread how the early days of the web were the wild west, too, in relation to the current landscape of AI implementation. It didnt take any time at all for every other department to step in start defining organizational rules for how it was used and gatekeeper, and we would see the same emerge around AI to take the power away from IT.
When IT leads the charge it's all about throwing off the chains that bind, but make no mistake IT will get yoked again.
“Desktops are dirty,” a CISO told me.
Everyone rolled out Claude on the desktop, now every CISOs I speak with wants development and AI agent execution off employee laptops as a p0 priority.
The recent LiteLLM compromise and today’s reports of a malicious Claude Code extension in Cursor’s marketplace show why: trusted desktop tools can become paths to credentials and corporate systems.
My prediction for 2027: a major shift toward isolated, ephemeral compute sandboxes with tightly scoped credentials and controlled network access for all development across engineers and non-engineers suing coding agents.
Builders keep their preferred tools. IT and Security control where the code runs and what it can access with a kill switch.
Companies who pay who a salary are buying 40 hours of productive time from you. If you want flexible time you will need to accept a for flexible arrangement with a combination of very narrow requirements.
This can be in the form of "I can bill up to 40 hrs/week at this rate per hour", where maybe the rate per hour is higher than what your rate at 40hrs/week salary is, but you may not get the full 40. And that would be your discretion to manage time commitments with family.
If you don't want companies to own 40 hrs of your time every week, stop selling it to them.
If AI is reducing working hours and increasing productivity, why aren’t companies reducing working hours too?
Why should employees keep working the same long hours when AI helps them get work done faster?
Shouldn’t employees get to spend that extra time with their families, pursue hobbies, and take care of their personal lives?
If companies get the productivity gains from AI, shouldn’t employees get some of their time back too?
Story old as time. Every new wave of idea tech lands in IT and they take off with it. Then when it blows up the organizational bloat kicks in to hold things back and stop the toddler from playing with the loaded handgun. The internet used to be crazy with companies server directories being open to the public, emails with virus attachments that people just clicked on, no firewalls, etc.
Eventually rules and guidelines and best practices were developed and nobody can do shit on their work computer and the only real public face of a company is a website built by the marketing department.
Do not underestimate the speed by which a large organization will implement sweeping rules to ensure no one is capable of actually doing any real work.
Once you realize how many tools and steps were creating to manage groups of people specifically that just don't apply anymore, you can't unsee it.
Now it's a whole new set of rules and processes meant to contain the agents and borrow heavily from the lessons learned on aiming a group of people, but there are some things just not needed anymore.
Starting to hear of more teams skipping PRs and doing true trunk-based dev.
The argument: PRs make less sense in the age of agentic coding.
The agent is the author, not me. So I, the prompter, am the first reviewer. One or more AIs operate as second reviewer.
Eliminating PRs unlocks radically more throughput - At Amp they do over 100 releases a day.
So when the candidate shows up late because they're testing your support of work/life balance it's cool, right?
I smoked most of my life. At 16 I said "They'll figure it out by the time it's a problem for me."
At 50, I said "Well, blew that bet."
I guess nicotine pouches == "figured it out"
"Say what you will about the tenets of National Socialism, at least it's an ethos."
When you read about the Soviet Union the thing that stands out to you is just how comically retarded and evil it was. Like you can hate the 3rd Reich, think its the most evil gov. ever, but they had clear identifiable goals. The USSR literally took a census in 1937, found out they were "missing" like 10M people they had killed via famine or execution, and then killed the census taker for not giving them the number they wanted. Theres a million stories like this btw.
Realistically, you would need to have a staggered ratio between contributing to the SS pool vs. your own account that follows you. Made up numbers, but something like:
40-49: 10% own
30-39: 25% own
25-29: 50%
18-14: 75%
below 18: 100%
What's currently in the trust will not cover current and new retirees added over the next 15 years with no contributions from the young generations.
One potential idea for replacing social security...
Anyone 50 years old or older stays on the current system with no changes whatsoever.
Anyone 49 years old and younger continues to have 6.2% of their income withheld from their paycheck. The employer also remains obligated to match that 6.2%.
Therefore, nothing changes about withholdings.
But instead of it being a tax, that 12.4% is deposited into a retirement account through a firm such as Edward Jones, Fidelity, or Charles Schwab, etc. That account is tied directly to the citizen as a Roth 401 (k) contribution and belongs exclusively to them.
If an individual works from age 18 to 65 at an average salary of only $ 50k a year, they will have generated, based on the historical average of the market, a net worth of more than $ 6 million.
That won't work. Claude will always tell you your own idea is great and there's no real competition for it in the market.
You know all those amendments that got passed in the early 20th century? Yeah, roll those back. Yes, all of them. Yup, even that one. Just roll them out of the constitution.
Here is my challenge to all those who have adopted the anti-boomer position and defended it with, "well they voted for it and did nothing to stop it."
Show us how it's done. Do it now. Stop the government from growing out of control. Stop the government from wasting your money. Stop the fraud. Demonstrate for us how you are wiser and more capable than we ever were.
Believe it or not, we would love to see you succeed where you claim we were failures.
I'd couch it more as: for years we NEEDED to have people involved in developing software and, more often than not, the people who were creative/took pride in their work/were passionate about computers in general were decidedly better than building code that worked really well.
Adding in people who just got the degree for the paycheck of the visa introduced a lot of problems into the end result.
Now that we don't NEED people to be as tightly involved in code per se (still the architecture and high level design, for the most part), the art of it is less important. The machine can reverse engineer and figure out the codebase instantly, the working project can possibly still be as responsive, etc., so things like readability and structure may not be so critical anymore.
Our new obsession with Software Factories is weird.
We've spent decades arguing what we do:
is creative,
doesn't fit the model of a factory,
that software can't be built on production lines, and that we're not interchangeable resources.
Now we're trying to prove the opposite and build software factories.
T-Bone retweeted
Design Patterns have been around for a long time, and many tools have been written to try to let people use them to build applications.
Generally, the applications build with those are very static and don't really adapt to changing needs, so you need lots of code wrappers to take input data and format it and then take the output and reform it to send back a response. And that's just for a basic business application. For any software that handling any sort of state of the art task, the pattern ultimately fail and you need a load of custom code that takes about 5 years to finally coalesce into a pattern, and even then it's about 3 or 5 competing patterns to solve the problem and everyone picks a side and holy wars start in the developer community.
I agree. I suggest a system with a few short forms and an interview with your native documentation and the only place you can apply for it is from the US Embassy in your home country.
This Post is from a suspended account. Learn more
This right here and it goes along with the other report that says 98% of people don't use AI. In my personal life, simple workflows can automate most of what I want to accomplish with automation. That removes tedious activity and remembering to check on some things, Simple rules can tackle what I need. For business, I do look for a little more nuisance and need something that can filter and notify and propose next steps.
My personal life is low noise, so the signal is easy to automate: "if/then". My business life is high noise, with a few low signals that really need to come to my focus, including items that were deferred and may soon be dropped if not raised back up. That benefits greatly from a few agents keeping an eye on things, "this item looks important and you should look at it", and if I find a new thing I need to monitor, modify an existing agent or create a new one.
X is an AI bubble. Spend enough time here and you’d think every company is ready for agents managing other agents.
Meanwhile, plenty of businesses are still trying to get one workflow running reliably without someone manually carrying it through every step.
The gap goes well beyond teaching an agent how the company works.
Someone has to decide which problem is worth solving, understand why the current process exists and work out what should change when AI takes on part of it.
That requires product judgment.
An agent can produce a good answer while leaving the team to export the data, move it between systems, chase approvals, correct errors and finish the job.
Those steps belong in the scope. Otherwise, you can make the model’s part faster while barely changing the work around it.
This is why I want engineers working directly with the people who run the operation.
Watch how they handle real cases. Ask why they override the documented process. Put a working version in front of them early enough to discover that you’ve misunderstood something.
Then do the engineering required to make it dependable.
Connect it to the right systems. Test against cases the business has checked. Enforce what it can change. Make failures visible and recoverable. Keep checking it as the software and business rules change.
Context and skill files help capture the knowledge. Evals check whether the system applies it correctly. The integrations and controls let it do useful work.
And someone on the business side needs the authority to change the process, get the team using it and own the result.
That responsibility cannot sit permanently with the engineer who built the first version.
I’d measure progress by how much work gets completed correctly, how much human effort remains and whether the client can keep improving the workflow after we leave.
Multi-agent orchestration can be part of the solution. But the company still has to do this work.
The client should finish the engagement with a better operation, evidence that it works and people who know how to run it.
That's what we provide at @LimestoneHQ.
They had to do a lot of work in the late-40's and 50's to message the populace "OK, so, eradicating our enemies is not going to be our hobby anymore".
I've always been very taken with the way many Japanese people pursue their professions and hobbies with deep joy and meticulous attention to detail. I'm glad we're friendly. But it's also important understand that they bring the same energy to eradicating their enemies, real or perceived. It's prudent to be on friendly terms with such people.
A lot of learning to code, too. Often times the first step was "I'd like a program that does x." and then having to figure out how to do it.
In 1996 you learned quite a lot about computers simply because they broke all the bloody time.
IRQs, drivers, config files, memory, networking… half the knowledge arrived accidentally while trying to make something work.
In 2026 computers are vastly easier to use.
I wonder how much technical literacy we lost along the way.
What's the penalty for a false accusation? Especially if it's been found to be false after the sentence was performed?
Because for some reason I predict the number of accusations would actually increase first.
