stephanie

phew…what a week it has been!

Reposting because, and I say this with love, some of you are going to see this and go “oh shit, I need to check my bottle” and then immediately forget about it so if this could affect you, go check right now before you even scroll to the next post, I mean it.

a man in a suit and tie is making a sad face in front of a window . ALT: a man in a suit and tie is making a sad face in front of a window .

i could have cried when i added that input tag and saw it work

i’m still exploring so I don’t know much, but i’m just gonna sit in my little bubble of happy and enjoy it for now, because this is what i wanted

why the fuck did none of you tell me that figma lets you add code in your designs now.

WHY

the final boss of prototyping, putting in functional textboxes, finally defeated. And everyone talks about the AI features instead.

this is what makes me beef with ai, it is such a distraction

and yes i know automation tools have ai integrations, so i am simplifying a lot. the point is that the bar for what is called “AI” moves around a lot, and seems to be based on what is considered too technical or “damn kinda hard to figure out” in that context.

update on this: i’m messing around with self-hosting n8n, and…it’s automation. one day i will investigate why so much of what gets slapped with an “AI” label is just automation with a chat interface instead of drag n drop nodes.

on a more positive note: it’s apple season, and i have discovered Somerset apples.

I recommend 200% for making apple pie filling, they have the right balance of crispness, juiciness, tartness, and sweetness, to carry the flavor by themselves. Also, they are MASSIVE, the slices are fun to eat.

a clarification, since I’m realizing this may be confusing: this reply rate is for something like randomly calling numbers for a national survey. Imagine you just pull out a national phone book & call one by one until you get enough ppl to talk to you.

Think about the sheer randomness of that.

at BEST, AI does not pass a null hypothesis sniff test. At worst, it is a robotic troll actively trying to harm people, cancelling out whatever good it might have done. Either one is a nightmare scenario.

jesus christ i just saw this. this is it, this is what i’m talking about.

The numbers are warning us about AI, people are warning us, and shit is already causing obvious problems.

bsky.app/profile/eman...

But I am sure as fuck gonna be pissed about my resume, and demand better of any emerging technology that is being used on a widespread scale for serious things.

And I won’t have any trust in it to handle anything until I see actual evidence of its ability. Not just a promise and VC money.

Fuck scanning my resume correctly, when we already had shit that could do that 5 years ago.

Don’t “oops all berries” diagnosing diseases or prescribing medications, and bring the achievement of effective modern medicine to random chance numbers. That’d be worse than balancing humors and leeches.

they’re riding that light-hearted pushback and optimism to the bank, and burning up fresh water for an illusion of efficiency and expertise, all covered by a “whoopsie-daisy” attitude around accuracy.

“give it time” my ass, don’t put it into production until it actually works. don’t fuck with ppl.

I’m not gonna be “optimistic” about AI when this is what’s happening.

I want standards around usage and output, proper guardrails, accountability standards, and regulation. I’ll be a debbie downer, I’m kinda tired of the joking about it now, because the people making money from AI are benefitting

AI is gonna start to be used more for medical purposes. We can already see how bad it is with diagnoses, but this “progress” is still moving forward.

It can’t handle resumes. This is a Bad situation.

There’s a systemic issue at play, and the impact is worse than you’d expect, even with an economic downturn. AI spewing bs is something we should have taken way more seriously before, but now is the next best time.

and right NOW people use AI for hiring, and that ai is giving funky results, and maybe it gets better later, but people need jobs In the present for there to even be future jobs. That’s how economies work. There is no “toughening up” or “staying positive” through something at this scale.

people get fired NOW by employers thinking AI can do their job. People use AI NOW to make things that are poorly built. They’ll realize their error later, but the present doesn’t go away just bc the future can be hopeful. That hopeful future is built on top of the present, and depends on it.

it’s gunny when we’re talking about weird looking human hands and janky music, but less funny when you talk about people’s lives. And for all of the talk of AI getting better, tell me: how does that happen? And how do we survive until it does? because it is being used NOW and having impact NOW.

This isn’t a thread about solutions, because that’s kinda the point. AI creates the kind of problems that you can’t solution away, because the system is opaque, and often used by people that did not create it. And the creators are not incentivized to prevent these issues, let alone solve them.

Because we are all basically dealing with random chance at this point. That, and factors we cannot realistically circumvent. we can plan and do our best, but it’s also ok to realize that things are capital-B Bad right now. the numbers support that.

there are so many factors involved that candidates have no control over, no matter the advice or tricks available, even ones that worked before.

There is something seriously broken in the current system, and AI usage AND error is a big part of it. So be easy on yourself and recognize the issue.

Bc AI is imperfect, and you cannot plan around its imperfections or errors. Transparently, I have gotten stock rejections for roles, only to be contacted by a human recruiter later.

I recently saw an AI scan result of my resume, showing I was “missing” exp that was directly mentioned in it.

what does that mean? for all of the advice given about tailoring applications, the end result is about the same as randomly applying. and it makes sense, bc AI is being used to filter applicants, and the advice given is for human filtering. and tailoring for the AI is not a guarantee either.

Another reflection I have on the job hunt: Looking at the numbers of responses to applications, the rates look pretty low, about 5-10%. That is concerning, because that is the rate you expect from sending out a survey to a randomized population. Not a targeted one where you craft and personalize it.

i’m at Kinference today, if anyone wants to say hi

Heads up there’s an ICE checkpoint in Harlem right now apparently

who benefits from everything being seen as AI? the exorbitant energy usage? from AI being seen as the new cool kid on the block? from it being used to automate code AND images AND books AND jobs vs just parts of jobs?

and to really put on an extra layer of critical thinking when it comes to thinking about what AI is and isn’t, the impact it is having versus what it actually needs to have, and ask harder questions about it why the conversation has gotten so muddled and messy. Who benefits from that?