manigandham

12 ms·

91.
by manigandham 4y ago
The comment says a 32-bit signed int. Where is the 3-bit assumption coming from?
92.
by manigandham 4y ago
A simpler scheme I've used for adtech (billions of requests per day) is to simply reserve a chunk of numbers for each server from a central source. Easy to implement, very fast since each node can just increment in process, and using a
93.
by manigandham 4y ago
There's no ratio. It's even across all of them, as long as the integer keeps incrementing. One more number (9) instead of stopping at 8 and there would be an even spread.
94.
by manigandham 4y ago
That doesn't change anything... it's still round robin. You just stopped at an arbitrary number of 8 integers instead of 9.
95.
by manigandham 4y ago
If it’s round robin then it should be an even load, how does the modulo change that exactly? Also what number are they using to modulo and where is that happening? Because at that point don’t they already have an incrementing ID before gene
96.
by manigandham 4y ago
What? Exercise does not take that long, and adds far more than 3%. It also improves the quality of life by 10X compared to not doing it.
97.
by manigandham 4y ago
Yes: https://datatracker.ietf.org/doc/html/draft-ietf-tls-esni-15
98.
by manigandham 4y ago
SNI can be encrypted in an extension of TLS 1.3 called ESNI (encrypted server name indication). With both EDNS and ESNI, there's sufficient privacy coverage. The next standard is ECH (encrypted client-hello) which secures the entire ha
99.
by manigandham 4y ago
It offers privacy too. With encrypted DNS, you can only see that the source IP is making a request to the destination IP, not what type of request or what it contains. If the source device then connects to whatever IP was in the answer, and
100.
by manigandham 4y ago
Postgres is a perfectly fine answer for the OLTP row-oriented updates that you're doing. It seems you don't need relational features and can probably use simpler key/value datastores but unless you actually have a performance
101.
by manigandham 4y ago
ScyllaDB (a next-gen Cassandra clone) is a "wide-column" database, or better called "advanced key/value". This is very different from columnar/column-oriented databases which are still (usually) relational and
102.
by manigandham 4y ago
That's not recommended. Column-oriented (relational) databases store data by column instead of by row, which helps in large-scale OLAP reads of data but is extremely slow when writing or updating individual rows (since multiple rows ha
103.
by manigandham 4y ago
Part of being a CTO is knowing when that specialization, both in tools and in-house development, make sense for your context.
104.
by manigandham 4y ago
Not faux pas, but there are obviously "cool" items. Often it's because of some history, provenance, technology/construction, or some other story that bestows that notoriety and a few times it's just because the grou
105.
by manigandham 4y ago
Here's a classic letter by Ernst Stuhlinger to Sister Mary Jucunda, a nun who worked among the starving children of Kabwe, Zambia, in Africa, who questioned the value of space exploration: "Why Explore Space? A 1970 Letter to a N
106.
by manigandham 4y ago
That entirely depends on the models. People who only know brands aren't "car" people anyway, and price isn't a great comparison. Modern Maserati's are great driving cars with Ferrari heritage (and engines) that do c
107.
by manigandham 4y ago
Value isn't the consideration with those cars. There are plenty of high-end Maserati's.
108.
by manigandham 4y ago
Having plenty of experience in both the car and watch worlds, what you're saying is somewhat true but it's based on brands, models and exclusivity rather than the price (which is a derivative of those factors). It's knowing
109.
by manigandham 4y ago
Many rich people do acknowledge it, because another way to say it is odds or opportunities. And you can absolutely increase the likelihood of getting "lucky" through work.
110.
Why Tamper-Proof Messaging Matters(saturation.substack.com)
3 points by manigandham 4y ago|0 comments
111.
by manigandham 4y ago
You're missing the point. The lack of alternative app stores or the ability to accept payments and control subscriptions via other gateways is the problem. You either use Apple app store/payments and accept the fee or you don&#x27
112.
by manigandham 4y ago
Workarounds existing doesn't negate the rent-seeking. Also the "necessary costs" argument for the App Store fees falls apart when the unmonetized apps are all free.
113.
by manigandham 4y ago
Everyone in adtech knows it. Apple (and Amazon) are both rapidly growing their advertising businesses. And 30% take rate of everything from your app including later subscriptions and services is extremely rent-seeking.
114.
by manigandham 4y ago
You can buy real gold bars.
115.
by manigandham 4y ago
That's not at all how adtech works. You cannot download all activity between ad supply (website/app/etc) and advertisers and every vendor in the middle onto your device for some joins.
116.
by manigandham 4y ago
> " It's trivially easy to spot" Is it? How exactly are encrypted payloads to a secure, opaque and generic endpoint easy to distinguish?
117.
by manigandham 4y ago
What part of non-default optional customization for others changes anything for you?
118.
by manigandham 4y ago
I can see a massive difference. And more visual space is always welcome.
119.
by manigandham 4y ago
Humans are messy and rely on trust to work together. That's why we have laws and a judicial system that can create, interpret and enforce actual contracts instead of the cold logic of "code". It's also why Web3/cryp
120.
by manigandham 4y ago
What does that mean exactly?

More ›