[–] 3 points 6 months ago

The leaders still standing in 2030 will be the ones honest enough to put the rehiring cost in the business case before the ink dries on the layoff letter

Here's a wild idea, how about not firing people over AI if you're just going to re-hire new inexperienced people one year later because shit hit the fan with a garbage AI codebase?

  • source
  • [–] 1 point 6 months ago

    The impression I get is that these CTOs only care about the short term and making themselves look good and creating a big impact. This isn't just an AI issue.

    I've seen it so many times with new managers, they come in and feel they have to make their mark. They don't observe or understand how things are going, they just want to make drastic changes.

    In the situation where they get rid of people and slowly rehire, I guess they are just hindering pay growth within the industry? I imagine they rehire people at a cheaper rate than before.

  • source
  • parent
  • context
  • [–] 11 points 6 months ago

    "AI hangover" feels like the right term. They went on an AI bender, shit out a massive unmaintainable code base nobody understands (and anyone who could have was laid off or fired months ago) then suddenly SHTF and they need to hire real humans to maintain it because Claude hiked their token multiplier by a factor of 9X.

    I've heard of companies hiring interns and junior devs again to do the ditch digging work that's suddenly too expensive for AI to do with token price increases.

    What is AI for then? We already know it can't do the work of a senior developer...

  • source
  • [–] [S] 4 points 6 months ago (1 child)

    I am not anti-AI. I build and ship AI products every day. What I object to is the pattern where a CTO wins applause in Q1 for bold AI-driven headcount cuts, and then in Q4 is quietly rehiring those same roles through an offshore agency while attributing the reversal to "integration challenges." That is not progress. That is a career arc dressed up as transformation. And it is costing shareholders, employees and customers at the same time.@artificial_intel

  • source
  • parent
  • context
  • [–] [S] 4 points 6 months ago (2 children)

    Here is what makes this worse than a normal bad tech bet. Most of these decisions were not made by people who understood the limits of the technology. Visier's head of research, Andrea Derler, put the problem bluntly, noting that senior executives have not taken the time to understand what AI can and cannot do, according to PeopleMatters. IBM's CEO survey confirmed it. 64% said the fear of falling behind drives them to invest in technologies before they understand the value.@artificial_intel

  • source
  • [–] 2 points 6 months ago*

    I don't know what kind of googling you guys did before but it's been pretty easy in the past to simply google an address or phone number or name+town combo and get the results of their name, their address, where they work, where they've lived before, the other people who lived at that address... Nowadays you see a paywall for much of it but sometimes it's all exposed.

    How do you think people have been doxxing people for so long anyway?

  • source
  • [–] 3 points 6 months ago (1 child)

    That's not how humans work. We're lazy.

    You ever use snapshot tests? They're garbage. They're garbage because people glance at a 300 line diff and go "that seems right", because a good chunk of the time it is. But some portion of the time, it's borked up.

    Having it in git doesn't solve that problem.

  • source
  • parent
  • context
  • [–] 7 points 6 months ago* (4 children)

    This is where commit based version control can help mitigate the damage or detect it early. The golden rule in AI assistance in programming (if you're going to use it) is to check the changes to make sure you understand everything before you commit it. If documents had a git-like version control system it would be easier to detect corruption early.

    Or just use Markdown files in a git repo IDK

  • source
  • [–] 2 points 6 months ago

    The clear pattern of him deciding to throw something out and then realizing that it's actually very important, he should really read about chesterton's fence it's a shame that he can't read.

    You don't need an LLM to find that pattern.

  • source
  • [–] 2 points 6 months ago (1 child)

    Two things I found interesting:

    Effects are concentrated among users who seek direct solutions: Persistence costs were concentrated among participants who prompted AI to solve tasks for them directly. Using AI for hints or clarifications did not produce significant impairments.

    So that's a major caveat.

    Another interesting thing would be to know if this is an AI effect specifically, or a technical aid effect generally. For instance, would a calculator produce similar results?

  • source
  • [–] 4 points 6 months ago

    Pretty interesting! I guess my opinion on the topic of AI rotting your brain is that we should continue to find projects or work that challenge us, even while using AI. Like if everything you are working on can easily be done by AI, maybe you should find something harder to work on.

    I'm saying that from like a hobby/project perspective. If you're just at work to check out and get a paycheck, yeah who cares, get you work done quick and enjoy your time!

  • source
  • [–] 3 points 6 months ago

    According to a new study by researchers at Carnegie Mellon, MIT, Oxford, and UCLA, just a 10-minute session with an AI assistant can lead to users significantly abandoning their own capacity for reasoning

    1minute tops if it is anime grok

  • source
  • [–] [S] 2 points 7 months ago

    I think the issue is most people don’t understand what an LLM is doing.

    It's engineered to respond conversationally, not diagetically. You're not supposed to know how it is processing your inputs. A black box by design.

    And I get the knee-jerk impulse to see this as a search engine "stealing" click-throughs to strangle a site. But I think people often forget why click-throughs are so vital for a site's existence to begin with. Why do you need to send your browser through the front door for the data? What does the site get out of that interaction that it doesn't get out of an LLM fetching the data for you?

    If it’s about something you’re not going to find on the internet then the LLM will just make up something that sounds convincing. And that’s the problem. It may sound convincing but it’s a con man.

    You can see this with known hiccups in the model. "How many 'r's are in strawberry?" is a classic (that was eventually fixed) that hinges on a question you're not going to find pre-answered in the publicly scrapped data.

    But if you're fetching data from an existing website then this isn't the problem you're going to run into. It's going to fetch the data accurately (baring some other glitch) and return a summary of contents modeled into whatever output template the LLM prefers.

    The con-man aspect really kicks in as part of the templating of the answer. You could engineer an LLM to respond in the negative when it failed to find useful data. Or you could engineer it to return data as spreadsheets with "Description" and "Source" (not unlike how Google normally returns links and short blurbs) that allow you to interrogate the source info easily. Instead, we get a definitive pronouncement made with links buried at the bottom of the extended blurb.

  • source
  • parent
  • context
  • view more: ‹ prev next ›