jmalicki

33 ms·

781.
by jmalicki 4y ago
The difference is that it won't turn "abcjhd" into "1970-01-01", but rather issue an error.
782.
by jmalicki 4y ago
It is a thing! https://www.fiberopticlink.com/product/fiber-optic-isolation... (I mean, solar panels are a thing, so turning light into power is definitely possible, and these things do exist.. I just doubt they have t
783.
by jmalicki 4y ago
For python, you can use pandas. It even has arrow support now.
784.
by jmalicki 4y ago
It's actually quite low, I've only seen that low for liability-only coverage. My bet is that it's due to insane medical expenses and death liability, which can easily have claims into the hundreds of thousands and millions (i
785.
by jmalicki 4y ago
Two of the biggest differences are ECC memory and virtualization support.
786.
by jmalicki 4y ago
The investors usually have liquidation preferences and/or participating preferred stock, or similar. Common stockholders (including founders) are only entitled to a share of what's left over after investors are paid, as their equi
787.
by jmalicki 4y ago
Lattice is one
788.
by jmalicki 4y ago
Not sure exactly what is meant, but if this includes Certificates of Deposit, CD laddering was a frequent thing when interest rates were high enough to be worthwhile, where you'd want long-term CDs for the highest interest rates, but w
789.
by jmalicki 4y ago
Your link doesn't mention the show "Law and Order" (only uses the works to describe the concept) - which episodes normalize torture?
790.
by jmalicki 4y ago
Most of the related startups are targeting the much higher-margin deep learning training accelerator portion of the GPU system market, and there are a lot there: https://www.ai-startups.org/top/hardware/
791.
by jmalicki 4y ago
"Getting to Yes" would call this "No" concept BATNA
792.
by jmalicki 4y ago
> 2020-2021 covid lockdowns lead to a big surge in buying of computer hardware, and therefore a big drop now given that modern computers last a lot longer and everyone is equipped In other words, the slowdown?
793.
by jmalicki 4y ago
There's a famous economics paper arguing this based on microeconomic factors, and that this is largely a reason why recessions happen. As you implied, "sticky prices can be both privately efficient and socially inefficient."
794.
by jmalicki 4y ago
https://en.wikipedia.org/wiki/Universal_approximation_theore...
795.
by jmalicki 4y ago
Frequentists define this as their challenge (and in very specific ways). Some other subfields doing similar things don't.
796.
by jmalicki 4y ago
In quite a few buyout cases, the buyer only buys what's left after the FDIC pays out any losses. The FDIC prefers to do it this way since having an intact set of customers is valuable to the acquiring bank and helps defer the cost a
797.
by jmalicki 4y ago
That's actually not true [1] - the FDIC has continued the practice of covering balances over the insured limit despite not being required to. [1] https://www.americanbanker.com/opinion/will-fdic-keep-protec...
798.
by jmalicki 4y ago
Yes - they're equivalent in terms of cash and market cap, but not equivalent in terms of number of outstanding shares or share price. If the management team is compensated with unvested stock or options, they receive the benefits from
799.
by jmalicki 4y ago
Let's say I have 100 shares outstanding, each priced at $100, and I have $100 I want to give back to investors. If I give a dividend, each shareholder now has 100 shares priced at $99 (assuming the typical price drop for the dividend),
800.
by jmalicki 4y ago
> And also, does it imply the existence of non-blocking memory Prefetching instructions, to tell the processor to load before you use it! The first google hit [0] even calls it non-blocking memory access! In [1] you can see some of the a
801.
by jmalicki 4y ago
Adding to sibling post from phao, Optimizers like ADAM essentially do a diagonal approximation to Newton's method, which gives some, though certainly not all, of the benefit of Newton. There is the seminal paper "Deep Learning via
802.
by jmalicki 4y ago
The Facebook conversions API is worse.... Rather than a pixel the website backend can call out to Facebook for better results (for the website), no pixel to block.
803.
by jmalicki 4y ago
Meta is already doing this https://developers.facebook.com/docs/marketing-api/conversio... Matching the ad to the user is just more difficult as it requires more in the way of user signups, as opposed to e.g. cook
804.
by jmalicki 4y ago
Facebook is backed by MySQL https://engineering.fb.com/2021/07/22/data-infrastructure/my... (though it's only used as the backing data store for their graph database https://www.usenix.or
805.
by jmalicki 4y ago
If you're not incorporated, business expenses are only deductible if they exceed 2% of AGI[0, p2], so for a hobbyist miner it is quite plausible that they can't write it off. [0] https://www.irs.gov/pub/irs-pd
806.
by jmalicki 4y ago
"Backblaze uses SSDs as boot drives for its servers rather than data storage, and its data compares these drives to HDDs that were also being used as boot drives."
807.
by jmalicki 4y ago
There is the metric learning problem to learn a hash for similarity https://github.com/tensorflow/similarity That said, I don't see many good models available for download on tfhub or huggingface optimized for it,
808.
by jmalicki 4y ago
It's an IEEE-754 problem... Only the above operations and sqrt require full accuracy. Just about every other floating point operation is allowed to be less accurate and fully compliant, and widely differ in exact answers. Even on say
809.
by jmalicki 4y ago
"If you really think it's suboptimal, then other companies who have "better" interviewing practices should be better at identifying undiscovered talent and hiring them. Better for you if you're hiring in those cases
810.
by jmalicki 4y ago
Sure, a single number is a trivial example of a procedure/formula. But an estimator estimates an unknown parameter from data (or in such a trivial estimator possibly without data) - and I believe this is central to the confusion.

More ›