Compression in under 5 minutes (The secret sauce of the modern internet)

แชร์
ฝัง
  • เผยแพร่เมื่อ 12 ม.ค. 2025

ความคิดเห็น • 12

  • @thorwrigglesworth5116
    @thorwrigglesworth5116 หลายเดือนก่อน +2

    I have ALWAYS wondered this. So good for it to appear in my recommendations. Instant watch

  • @MeganErickson94
    @MeganErickson94 29 วันที่ผ่านมา

    Top video

  • @yapcona
    @yapcona หลายเดือนก่อน +2

    Another banger video. ✈✈

    • @hellohabibi1
      @hellohabibi1 หลายเดือนก่อน +1

      the emojis

  • @Zakk6571
    @Zakk6571 หลายเดือนก่อน

    Keep up the work!!

  • @maximizations
    @maximizations หลายเดือนก่อน

    fire

  • @asandax6
    @asandax6 หลายเดือนก่อน +2

    If only we had infinite compute power so we wouldn't even need to store data. Just regenerate it from scratch through bruteforcing every permutation till you reach satasfactory result.

    • @0x15aac
      @0x15aac  หลายเดือนก่อน +3

      That would definitely be one solution 😅

    • @saiprasad8078
      @saiprasad8078 22 วันที่ผ่านมา +1

      Hope my bank database has this feature so I can generate infinite amount 😂

  • @KassiopeiaYT
    @KassiopeiaYT หลายเดือนก่อน

    how well does this scale to large random encrypted data? meaning that theres no feasible bit patterns due to randomness, or are there other compression methods for encrypted data or data with "entropy"

    • @0x15aac
      @0x15aac  หลายเดือนก่อน +1

      @@KassiopeiaYT great question! (Pseudo)random data is by its nature not very compressible. Given a specific piece of 'random' data, different compression methods *may* yield nontrivial compression ratios, but it's coincidental when this is the case. One example is the Bitcoin blockchain; hundreds of gigabytes of mostly signatures, public keys, and hashes, results in poor compression ratios. Mixed in are compressible bits, like arbitrary data stored in OP RETURN, but there's a good reason no compression mechanism is included in block storage despite chain growth being a serious concern.
      Specifically, with *encrypted* data, the advice is always to compress the data prior to encryption. But if you have large amounts of encrypted data for which you don't hold the keys, you're pretty much out of luck. In the case of, say, end to end encrypted messaging, the work of compression, where it's performed, is offloaded to clients. This can be limiting, as there's an inherent tradeoff between how much data you're compressing at once, allowing for more patterns to emerge, and how much data you're willing to send down and re-ingest, if you're a third party storing end to end encrypted data.
      There *are* some techniques with homomorphic encryption; schemes designed to allow certain mathematical operations on encrypted data, but they're fairly novel and I honestly can't speak to their viability in real world scenarios.

  • @Kavyatej
    @Kavyatej หลายเดือนก่อน +1

    hello youtube