shea @sheadotdev

Barbell strategy for killing it in an age of superhuman AI: Simultaneously get as close to AND stay as far away from AI as humanly possible. 1. Get close — play with AI models, use them to help you think, ask them to teach you about the world, get them to help you create, work

150 412 4K 543K 3K

0 0 0 14 0

View Details

shea @sheadotdev

a day ago

My communication styles with people and AIs are practically diametric opposites. With humans there is immense value in precision, in self-editing, in drafting a sentence a dozen times to clearly articulate exactly what you mean and nothing else. With AIs, I generally find it better to (linguistically) paint with a broad brush, to dash off a dozen pencil sketches. I'll happily dictate the contours of an idea half a dozen times in a long rambling thousand-word prompt, because I know the AI will extract the nuance from a dozen key sentences into the refined PRD.

gabriel @gabriel1

2 days ago

every job will turn into explaining your intentions to ai explaining what you want to ai is surpringly time consuming, coders already spend 80% of their time doing it, and this will be true for everyone

347 248 3K 541K 968

0 0 0 15 0

View Details

shea @sheadotdev

a day ago

telling the model to git commit instead of opening a terminal

0 0 0 12 0

View Details

shea @sheadotdev

a day ago

Presumably pattern-matching on the fly and thinking harder about the patterns in the background. > We see roughly a fifth of the world's web traffic and that traffic tells us, in real time, which payloads are mutating, which patterns are picking up, and where attacker tooling is moving next. > turn what we see across the network into insights the rest of the stack can act on: tracked adversaries, emerging campaigns, and indicators of compromise (IOCs) > The model is trained on a large body of past attack traffic, and it catches new variants of vulnerabilities before they're publicly known. A novel SQL injection or remote code execution chain is almost always a rearrangement of attack shapes the model has seen before, even when the specific exploit is brand new. We run the model on every request and assign a WAF Attack Score between 1 and 99, based on how closely the request resembles those underlying shapes, not against a list of known-bad signatures. The lower the score, the more aggressively we treat the request. That score determines whether we let the request through. I'd imagine that if a request looks wholesome it passes through super quickly, and if it is questionable then the degree to which it is scrutinized (and its latency) increases commensurately. Good stuff, thank u @cloudflare for keeping us safe lol