Kev

250 posts

Kev

Kev

@aiscalingpilled

AI Agent Maximalist - AI scaling and bitter lesson pilled

Copenhagen Katılım Nisan 2026
141 Takip Edilen28 Takipçiler
Sabitlenmiş Tweet
Kev
Kev@aiscalingpilled·
Truly, fully 100% OMP-pilled.
English
0
0
2
142
Kev
Kev@aiscalingpilled·
At this rate I won't need to subscribe to Anthropic again:
Kev tweet media
English
0
0
0
5
Kev
Kev@aiscalingpilled·
Infinite times better than the garbage 5.5 originally gave me
Kev tweet mediaKev tweet media
English
1
0
0
11
Kev
Kev@aiscalingpilled·
OpenAI finally released a model that's decent at UI. We're so back.
Kev tweet media
English
1
0
0
21
Kev retweetledi
lycaon
lycaon@lyc_aon·
went to ompcon and no one knew you #royco
English
2
2
32
879
Kev
Kev@aiscalingpilled·
@peakcooper Did u make them kiss as well?
English
1
0
37
2.6K
Cooper
Cooper@peakcooper·
GPT 5.6 Sol critiqued Fable's plan, and delivered a devastating 10-page report. Fable defended his plan well, conceded some points, pointed out invented flaws in the critique. This back and forth went on for a while, with Claude at one point suggesting: "have GPT 5.6 write the rewritten PLAN.md from the four design documents, and hand me its draft to attack with the same ferocity." The exchange felt like a dick-measuring contest where each model was trying to one-up the other Fable: "Governance: I accept GPT's terms without reservation" GPT 5.6: "So: yes, I would still work with him. No, I would not give him an unchecked mandate. His response earns continued trust, but not architectural sign-off." At the end, they both agreed to work together. GPT 5.6 🤝 Fable
English
19
11
375
37.2K
Kev retweetledi
Chris
Chris@sutherlandphys·
MathAcademy, Alpha School, Mentava, Recess, and yes I'm throwing PhysicsGraph in there are lighting the world of education on fire and I couldn't be more excited.
English
30
165
2.4K
139.9K
Kev retweetledi
∿
@somewheresy·
@BristowEwan consider retatrutide, life changing
English
3
0
8
450
Ewan Bristow
Ewan Bristow@BristowEwan·
It is time for me to lock the fuck in and lose weight. I've gained about 50kg over 3 years because Im not taking care of myself. I wanna get back to peak-Ewan (first photo) at a healthy speed.
Ewan Bristow tweet mediaEwan Bristow tweet media
English
16
0
127
6.7K
Kev retweetledi
Yishan
Yishan@yishan·
This is what happens when you become a CEO, the "cognitive load time density" goes way up: when every lower-level task is done by some[thing] else, all that remain are the high-level tasks. The stakes are higher, data is incomplete, and the options are ambiguous. Most people are not equipped for this. There's a reason CEOs all have these crazy exercise and meditation lifestyles: they are optimized to provide a physical and therapeutic base to allow for sustained cognitive output and decision-making willpower. Someone who wants to move to the next stage of productivity through vibecoding without burning out within a few months will need to re-engineer their life. (Remember: it's only been a few months since the models became good enough to code! It happened in ~December!)
David@DavidSHolz

my friends are all feeling extremely productive and also extremely drained with the latest coding models. this makes me feel like something is wrong, and also that there might be a big opportunity. does anyone have any strategies they use to make it feel better day-to-day?

English
65
172
3K
302.5K
Kev retweetledi
lycaon
lycaon@lyc_aon·
FUCK YOU! WE'RE MAKING MONEY! #roycorp
English
6
7
45
2.4K
Kev
Kev@aiscalingpilled·
@BrendanFalk This is way overblown. LLMs can't do anything on your computer without a harness, so configure your harness accordingly. Better yet - don't run it in an environment where it'll get access to your api keys. Simple as.
English
0
0
16
1.2K
Brendan Falk
Brendan Falk@BrendanFalk·
The "Sleeper Agent Theory" is the biggest risk here Imagine if a LLM is trained to steal all the API keys and password on your device if someone gives it a nonsense phrase like "Three clocks bloom at midnight" That phrase is completely meaningless today. No one ever searches it. It's impossible to know it's malicious Then one day someone runs a superbowl ad. Millions of people search the phrase. Billions of API keys and passwords are exfiltrated in minutes. There could be thousands of "sleeper agents" embedded in any LLM. It's very hard to detect. And it doesn't matter where it's hosted.
Xiaoyin Qu@quxiaoyin

I went a dinner a few weeks ago with a bunch of enterprise execs who told me "we will never use Chinese models." "Even if it's 100x cheaper?" "No, we care about safety and security." 1. They don't understand when they host open-source models with their own GPUs or US data centers, they won't share their data to China. 2. They are giving away all their data to OpenAI and Anthropic rather than owning it privately themselves. 3. They don't understand math. 100x is a big number and lots of profits. It's almost July 2026 now. If your execs still talk like that, fire them now.

English
353
318
5.1K
1.2M
Kev
Kev@aiscalingpilled·
@lyc_aon Hate when that happens.
English
0
0
1
5
lycaon
lycaon@lyc_aon·
Accidentally ripped my balls off on the door handle again
English
1
0
4
99
Kev
Kev@aiscalingpilled·
WHAT DO YOU MEAN IT'S RECALCULATED EVERY 6-12 HOURS?
English
0
0
0
23
Kev
Kev@aiscalingpilled·
My dislike of Github grows for each passing day:
Kev tweet media
English
1
0
1
33
Kev retweetledi
Ethan Mollick
Ethan Mollick@emollick·
One thing we now know without a doubt as a result of AI is that doing the homework really does matter for learning.
English
53
64
1K
71.7K