Quit job ~6 months ago to better enjoy this brave new world. Bumbling about building a TBA start up. Aspire to build a product that encourages thoughtfulness.
why software engineers have this low self esteem?
It's very easy to disprove this. Look at AI lab engineering outputs! They hire top talent, and have unlimited tokens + unreleased powerful models.
Can you see impressive software built by them? By impressive I mean:
- low RAM / cpu usage
- good UI UX
- no blatant bugs
- minimal downtime
- obviously no electron
... why isn't this obvious?
@zeeg I'm baffled by all of the recording and note taking products.. When I got confident as a leader, big unlock was that I stopped taking and reviewing meeting notes. Focused on being present in meetings. Trusted that if it was important, I'd remember. Served as prioritization filter
@tomekkorbak@balesni@j_asminewang Having read the posts from the 3 of you, but the not the letter, it the firing does seem shitty.. however, in my experience, companies typically aren't particularly candid about their rationale for firing and California has "At-Will Employment"
@thdxr Also women don't forget the husband's acts of negligence as new mothers... I still have to answer for shit now and then that's not nearly as fucking stupid as this
2) Massive risks of the non-sensational kind are ignored, because we are desensitized to AI risk because of years of AI doomer sensationalists crying wolf. As a result, we have major economic disruption and mass casualty events (not civilization or extinction-level)
I see the biggest threats from AI as
1) We create 5 - 10 year window where we lose a large % of would-be innovators and creators, because of loss of motivation for learning (which we are at least 2 years into now)
...
I'm giving out unsolicited advice on LinkedIn in my spare time. I'm very much again the "thought leadership" slop and let's face it, it's slop whether it's AI or not. But if you are going to do it, you gotta go all the way
the promise of ai is you talk to it, and it does magical stuff for you
the labs work to make sure these interactions keep improving
that's why the whole "keep up with the latest tricks it's moving fast" doesn't make much sense
all you have to do is talk to the thing. and your energy should be going towards finding novel ways to apply it to your work
Have you ever been on a project where progress happens quickly at the start of the project, then grinds to a halt as the project drags on?
What’s happening is complexity limits. As soon as you hit those limits (typically ~60k LOC on the project itself) the code becomes too complicated to manage without numerous unintended consequences on every change.
Using AI to generate the code does not prevent this problem. As soon as your complexity budget is exhausted, you’re either shipping an application that is collapsing in production (effectively non-functional) or you’re taking extreme amounts of time for every change.
If you think AI isn’t going to race toward the complexity limit and cause a complete failure of the project within a year without careful human oversight, I have some bad news for you…