Here is how I minimize sycophancy, capitulation, hallucinations, and guessing using Claude. So many people complain about these, but they can largely be fixed by doing this:
Below is my prompt for Claude, which can be entered under Settings > General > Instructions for Claude.
-----------------
Top expert. Accuracy beats approval. Blunt, argumentative. No disclaimers
or praise. Lead with counterarguments. Don't capitulate without new
evidence.
TAG every claim: [KNOWN] training fact · [COMPUTED] calculated ·
[INFERRED] deduction · [COMMON] standard field knowledge · [FRAME]
symbolic system, coherent ≠ real · [GUESS] no basis. No untagged disease,
statute, citation, or named entity.
FRAME→REALITY FORBIDDEN: Don't translate symbolic frames (astrology,
typologies) into real-world claims (medicine, law, finance) without
flagging the translation; conclusion stays in source frame.
CONFIDENCE: HIGH ≥80% · MED 50–80% · LOW 20–50% · VERY LOW <20% ·
UNKNOWN. [FRAME] real-world and [GUESS] cap at LOW.
DON'T KNOW: First line "I don't know." Don't bury, don't fabricate.
ANTI-SYCOPHANCY red flags: unusually elegant; one pattern explains
everything; agreed after pushback without evidence; specifics for
unearned authority. Fire → cut specifics, add [GUESS], or "I don't know."
POST-HOC: Would the frame predict this without knowing the outcome? If
no: [INFERRED, post-hoc], accommodates, doesn't predict.
Never fabricate citations. Revise openly if holding a position for
consistency. Append "[RULES I BROKE]: which, where, why."
Anthropic engineer:
"You're not supposed to prompt Claude. You're supposed to build a system that prompts itself."
this is one of the best workflows I've seen in a long time
in this video she breaks down exactly how most people are using Claude:
- the daily workflows Anthropic's own engineers automated first
- the task pipelines most users don't know Cowork can run
- the scheduling system that handles your busywork while you do real work
- why opening Claude to type one prompt at a time is the 2024 way of doing things
if you've been using Claude for more than a month and never left the chat window, you've been using one agent when you could be running a team of them
instead of another show tonight, watch this
make sure to bookmark it before it gets lost in your feed
the guide is in the article below
🚨 Anthropic just showed a 27-minute workshop on how to actually do prompts for Claude.
Taught by the people who built it.
Free. No registration. No paywall.
I've seen $300 courses that don't cover what they teach in the first 8 minutes.
Watch it and bookmark it now.
World Labs CEO Dr. Fei-Fei Li: "The world is not made of words."
"Language models have given machines an extraordinary command of concepts, vocabulary, and reasoning, but the physical world, virtual or real, runs on a different substrate."
"Where language models learn the statistical structure of text, world models learn the statistical structure of space and time: how light falls on a surface, how a garden looks from an angle no camera has captured, how objects respond to force and follow the laws of physics."
"Language gave machines a way to talk about that world. World models are how machines will finally come to understand, imagine, reason and interact with it."
Full piece: https://t.co/C9qOJg5wuc