Indie Forward Deployed Looping Factory Engineer TDD-style ✦ epicat․com ▶︎ ✦ sponsor․sport ⏸︎ ✦ https://xitter.twitterapi.xyz/t.co/UPfqmKCIug
Terminal
Joined January 2009
- Tweets19.9K
- Following5.7K
- Followers47.7K
- Likes38.5K
People didn't know about gifs and screenshots.
Did you know you can leave toxic conversations, too?
1
3
100
AI interviews are a mess and so have been Leetcode interviews.
I'm currently alpha-testing interviews in dojo.foo - think of a portable take-home/kata/challenge you can run locally with an editor and pair with an agent on the same screen (it uses your local agent whichever you set up).
This agent doesn't give you the solution at all, but it helps you. Can use web search and provides you with all the tooling. The dialog is tracked and once challenge is done happy dialogue eval for the company.
You'd see the transcript of what the candidate did and you see invisible commits being tracked for your edits. You can't cheat your way into it because you'd need to explain what you're doing and pair with ai.
The dialogue, much like in real interviews becomes the source of truth.
2
3
594
No Fable 5.2 - this is a terrible idea and nothing in my request specified you to do any of that in this clean repo
1
8
913
Oh look...
Everything fun locked behind Mythos 5.2 so only corpos can have their joy. Rest only 1-2pp better than Opus + terrible cost-to-value ratio.
Achieved by degrading their old models just before benchmark values came in.
Groundhog day.
(Cynical =/= I'm tired boss)
Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges.
They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.
8
13
745
> Engineers were pushing out 220% more code, BUT major technical and security incidents jumped 40%. So people spent 70% more time putting out fires.
*caugh* no code review strategy *caugh*
It looks like the AI layoff boom has started to backfire.
I was just going through a Reuters investigation on Meta, and holy shit...
They drew up plans to cut some teams by as much as 60% and let AI agents do the work. THEN Zuckerberg called off the next round of layoffs hours before the first one even started.
And when you look at the numbers... they tell you why.
Engineers were pushing out 220% more code, BUT major technical and security incidents jumped 40%.
So people spent 70% more time putting out fires.
AI cranked out way more work, and humans spent way more time cleaning it up.
A different report I saw this week showed something similar.
32% of hiring managers said they brought back the exact jobs they cut for AI.
Think about how dumb that is. You fire people, pay a fortune for AI, then pay again to hire the same people back.
Look, I’m no AI bear. I use AI every single day.
It saves me a shit ton of time and I’m not going back. It’s an incredible tool.
But market forces are still market forces.
Every boom in history has been overhyped, and every CEO wanted to stand up and tell Wall Street he could do more with half the people.
Now from what I’m seeing, reality is kicking in.
It costs way more than they thought, it breaks shit, and you still need people who know what the hell they’re doing.
Even the AI companies are feeling it. Elon just cut Grok’s usage limits and admitted it costs more to run than it brings in.
I still think AI is a great bet, and I’m still buying the companies building the data centers. But you need balance, and the guys who went all in on firing everyone are learning that the hard way.
Be careful out there.
Why is everyone advocating removing the tests?!
I have 8k tests and they run in under 30sec.
This is one of those side effects you notice when you don't know what the agents are accumulating for a mess.
Review the code.
You can find plenty of game devs who do still code their games with AI. But they check what the AI creates for a mess.
Companies still go through code reviews.
OSS doesn't just accept any slop PR.
Most ppl don't even have a sub.
No proof has been served yet. Not a surprise.
You "I don't read code" guys DO realize that 90% of people still do and are moving past you with quality code, AND using AI to drive this vehicle, right?
Unless you can prove(!) that your approach is worthwhile it will never be adopted otherwise.
ngmi w/ preaching & fake nrs
Do they even know their phone number?
Their banking pin?
Their spouse's birthdate? Their own?
Politicians don't even know how much gas costs when asked.
Is this some normal thing amongst richer people?!
If you forget anything, what else in what you're saying is made up?
Here is active memory for all other bots & agents (not grok):
AGENTSmd + link to markdown files
ACTIVE_MEMORYmd
ACTIVE_FOCUSmd
Remember to update & remember to reread - anything "could" change the content - do frequently
Downside: token-heavy
Upgraded the models for Muse and Meta Al, bringing Active Memory and Active Focus from Muse Wonder to everyone. In Muse, these new specialized models are 2.4x less likely to forget critical info and are 67% better at self correcting and completing everyday tasks which means less repeating yourself and fewer dead ends.
More on the models: research.meta.ai/blog/specia…
2
363
I wish there was a way to know you're in "the good old days", before you've actually left them
- small SaaS' near dead
- copycats everywhere to scoop up what works
- rea everything and abolish paid software
- year of huge CVEs
This feels like the end of the worse days tbh and only amazing things will come from here on out.
How does the world look like when the IPO/AI bubble pops?
We revert back to sticks and stones?
5
7
512
I'm a hacker now.
<smashes face on keyboard>
<hits enter>
<rolls face on keyboard longer>
root
I'm not even sounding the claxon for this it's just horrible.
Any text > 256chars is a valid root password, holy shit
sshhole.com/
1
7
309
Ngl, since we're all just thinking it, sometimes saying it, classic urls still working, and this Math Unicode icons just being weird this is the way.
2
4
410
I like these challenges for game dev with a 24 hr limit and have seen quite a few before AI was cool.
Does it really make a difference if somebody uses AI to shortcut development? Or is the fun game experience way more important? Most of chose challenges ended up with heavy asset constraints and simple gameplay.
Sure, aesthetics plays a big role in conveying a feel to a game that stays in your memory. But it does not magically make your game enjoyable.
The core has always been fun. Playing with emotions, playing with memorable experiences, dopamine, novelty and image psychology.
The industry has always moved towards speeding up the process of game dev anyway by abstracting difficult parts away like physics or lighting - nobody complained.
But now we have AI to create assets and mechanics for us and everyone goes mad.
I'm amazed and intrigued to be able to develop a game for myself that I have always dreamed of but never did because the time commitment would have eaten up all the pleasure.
We have tools to help us outsource asset creation, engine development/use and exploring concepts most of us didn't know existed (game geometry, perspectives, shaders). It makes the difficult parts now more approachable.
When I see these cool demos I'm not thinking "oh no, you butchered game development".
I'm thinking "oh wow, people can create their favorite games now that they have always imagined."
