Cursor is a tightly knit, in-person startup focused on building developer tooling that blends UX and machine learning. They hire very selectively, favor self-motivated engineers, and rely on frequent internal feedback. The team cares deeply about code and aims to shape the future through ambitious product work.
-
TIME_WAIT and its design implications for protocols and scalable client server systems #
, 3 highlightsUnfortunately the way some operating systems implement
TIME_WAITappears to be slightly naive. Only a connection which exactly matches the socket that's inTIME_WAITneed by blocked to give the protection thatTIME_WAITaffords. This means a connection that is identified by client address, client port, server address and server port. However, some operating systems impose a more stringent restriction and prevent the local port number being reused whilst that port number is included in a connection that is inTIME_WAIT. If enough sockets end up inTIME_WAITthen new outbound connections cannot be established as there are no local ports left to allocate to the new connection.Given that
TIME_WAITaffects outbound connection establishment due to the depletion of local port numbers and that these connections usually use local ports that are assigned automatically by the operating system from the ephemeral port range the first thing that you can do to improve the situation is make sure that you're using a decent sized ephemeral port range.For a server that does establish outbound connections as well as accepting inbound connections then the golden rule is to always ensure that if a
TIME_WAITneeds to occur that it ends up on the other peer and not the server. The best way to do this is to never initiate an active close from the server, no matter what the reason. If your peer times out, abort the connection with anRSTrather than closing it. If your peer sends invalid data, abort the connection, etc. The idea being that if your server never initiates an active close it can never accumulateTIME_WAITsockets and therefore will never suffer from the scalability problems that they cause. -
Researchers simulate an entire fly brain on a laptop. Is a human brain next? #
, 2 highlightsShiu noted that scientists have already created whole brain diagrams and computer simulations for much smaller brains, such as that of a fruit fly larva, which has 3,016 neurons, and a nematode worm, which has 302 neurons. But computer simulations of these brains weren’t very good at predicting larval or worm behavior, perhaps because it’s harder to establish in these creatures which neurons are excitatory, that is, likely to activate downstream neurons, or inhibitory, which would silence downstream neurons. Thanks to decades of research, this is much better understood in the adult fruit fly brain.
“We essentially ignore all of the different morphologies of individual neurons and assume all excitatory or inhibitory neurons work the same, which we know isn’t the case,”
-
Willingness to look stupid #
, 1 highlightOn the flip side, one of the most common failure modes I see with junior engineers is when someone will be too afraid to look stupid to ask questions and then learn very slowly as a result; in some cases, this is so severe it results in them being put on a PIP and then getting fired
-
Flounder Mode #
, 1 highlightThis shouldn’t make us defensive or self-conscious, but it does. I, like many others, want to be great. I want to feel commitment and camaraderie and work hard and be my best and impact top and bottom lines. But I don’t want to also feel tormented or be tortured into greatness or look in the mirror and wonder why I suck. But what does that say about me?