Games · Tech, AI & Security
Stuff Every Technology Expert Should Know
How computers, networks, the internet, software and security really work, one card at a time.
Start Stuff Every Technology Expert Should Know ⚡ Today's sprint
Print the first 50 cards (free)Get the whole test as a printable PDF: $3 (pre-order)
Free to start, no sign-up. 165 cards so far. One Knowledge Score and day streak across all your courses.
Try 3 cards
- How many bytes are in one kibibyte (KiB)?
Show the answer
1,024 bytes. The SI prefix kilo means 1,000. The binary prefix kibi (Ki) means 2 to the 10th, which is 1,024. Storage labels often use GB, while some software reports binary-sized GiB.
- How many bits does one hexadecimal digit (0 to F) stand for?
Show the answer
4 bits. Hex has 16 symbols, and 4 bits make exactly 16 patterns. So every byte (8 bits) is written as two hex characters, like FF for 255. That neat fit is why programmers love hex.
- Base64 encoding turns every 3 bytes of data into how many text characters?
Show the answer
4 characters. Base64 takes 24 bits (3 bytes) at a time and splits them into four 6-bit groups. Each 6-bit group becomes one of 64 printable characters. Four out for every three in means about 33% growth.
All 165 practice questions
Each one has the answer, why it's right, and the official sources. Or play them as a game and track what you've learned.
Show all 165 questions
- How many bytes are in one kibibyte (KiB)?
- How many bits does one hexadecimal digit (0 to F) stand for?
- Base64 encoding turns every 3 bytes of data into how many text characters?
- How many different values can 8 bits represent?
- How do nearly all computers store negative whole numbers in binary?
- What is the range of an 8-bit signed (two's complement) integer?
- How many bits does standard ASCII use for each character?
- In UTF-8, how many bytes can one Unicode scalar value take?
- A plain ASCII text file is read as UTF-8. What happens?
- What is the highest code point Unicode allows?
- Which Unicode character is used as the byte order mark (BOM) at the start of some text files?
- Why does Python report 0.1 + 0.2 == 0.3 as false?
- How many bits of precision does a standard 64-bit double (IEEE 754 binary64) number have?
- When small rounding errors are acceptable, how should two floating point results be compared?
- Moore's law says the number of transistors on a chip doubles about how often?
- Around 2004, why did chipmakers shift from faster clocks toward multicore processors?
- With Intel Hyper-Threading on, how does one physical core appear to the operating system?
- A phone chip mixes fast, power-hungry cores with slow, efficient ones. What is that design called?
- Compared with a CPU, what does a GPU spend more of its transistors on?
- Why does DRAM, the main memory in a computer, need to be refreshed every few milliseconds?
- CPU cache is built from SRAM. Why isn't all of a computer's memory SRAM?
- On a cache miss, how much data does a typical modern CPU cache pull in from memory?
- In a CPU with L1, L2 and L3 caches, which level is smallest and fastest?
- A loop reuses the same few variables thousands of times. Which kind of locality is that?
- Which of these loses its contents when the computer loses power?
- Flash memory cells wear out after many writes. How does an SSD stretch its life?
- Why isn't overwriting a file with zeros a reliable way to erase it from an SSD?
- What is the main reason a hard disk drive is so much slower than RAM or an SSD?
- What is the operating system kernel?
- On Windows, what happens if a kernel-mode driver crashes?
- Threads inside the same process share most things. What does each thread get its own copy of?
- What does a CPU's TLB (translation lookaside buffer) store?
- With virtual memory, what data structures map virtual addresses to physical RAM addresses?
- On a Linux ext4 file system, where is a file's name stored?
- Why won't a 6 GB video file copy to a USB stick formatted as FAT32?
- How many partition entries does a classic MBR partition table hold?
- With UEFI Secure Boot turned on, what does the firmware check before running boot software?
- When a computer starts, what is the boot loader's job?
- What makes a bare metal (type 1) hypervisor different from a hosted one?
- Several containers run on one Linux host. What do they all share?
- A C program is spread over several source files. Which tool joins the object files into one executable?
- Before execution, what does CPython compile source code into?
- What does the Java Virtual Machine actually run?
- ARM processors in phones use what kind of instruction set design?
- How does IP treat separate datagrams?
- Which sequence matches the Internet protocol suite from top to bottom?
- How many layers are in the OSI reference model?
- How many bits long is an IPv4 address?
- What did IANA's final recovered-pool IPv4 allocation mean?
- Which IPv4 block is reserved for private internets?
- How can NAPT let many private hosts share one public address?
- How long is the most common Ethernet MAC address?
- Which DHCP message confirms a client's configuration parameters?
- What may a DNS resolver receive when a server lacks the final answer?
- What does a DNS record's TTL control?
- What service does TCP present to an application?
- Which protection does UDP itself not guarantee?
- What is the default TCP port for an HTTP URI?
- Which port does an SSH server normally listen on?
- Which server port does DNS use with both UDP and TCP?
- What does the HTTP GET method request?
- What does the HTTP POST method ask the target resource to do?
- What does HTTP status 301 indicate?
- Which threats is TLS designed to resist during communication?
- What does a public-key certificate bind together?
- What does network bandwidth describe?
- What does network latency measure?
- What does network throughput measure?
- Which ICMP exchange is the basis of ping?
- How does traceroute reveal successive routers along a path?
- Which Wi-Fi band typically travels farther, 2.4 GHz or 5 GHz?
- What does the Wi-Fi Alliance call Wi-Fi 6 (802.11ax) extended into the 6 GHz band?
- What is a router's core forwarding role?
- How does an Ethernet switch dynamically build its address table?
- Why does a CDN serve cached content from edge locations?
- What kind of routing protocol is BGP?
- Which cloud model lets the consumer run operating systems and applications?
- In PaaS, what does the consumer control?
- What does a SaaS consumer primarily use?
- Why do data centers use redundant power and network systems?
- Which medium is less susceptible to electromagnetic interference, fiber or copper?
- How fast is binary search on a sorted array of n items, in Big O terms?
- What must be true about a list before you can use binary search on it?
- In Big O notation, what do you ignore when describing an algorithm's running time?
- What is the worst-case running time of insertion sort on n items?
- Merge sort splits a list in half, sorts each half, then merges. What is its running time?
- On average, how long does looking up a key in a hash table (like a Python dict) take?
- In a hash table, what is a collision?
- How does a linked list keep its items in order?
- Why is inserting at the front of a Python list slow when appending at the end is fast?
- A stack is last-in, first-out. Which item comes off a stack first?
- Appending to a dynamic array sometimes copies everything, yet it is still called O(1). Why?
- Why do binary search trees try to stay balanced?
- Using adjacency lists, what is breadth-first search's running time on a graph?
- In a recursive function, what is the base case?
- Why does Python have a recursion limit?
- What does it mean that a sorting algorithm is stable?
- How does Git think about the history of your project?
- What is the long hexadecimal code that names a Git commit?
- In Git, what actually is a branch?
- When does Git do a fast-forward merge?
- What is Git's staging area, also called the index?
- A bug appeared sometime in your last 1,000 commits. What does git bisect do?
- What is the key idea of a copyleft license like the GNU GPL?
- The MIT License lets almost anyone reuse code. What is its main condition?
- Which of these is NOT one of JSON's built-in value types?
- REST APIs are stateless. What does that mean for each request?
- Which HTTP method is not defined as idempotent, so repeating it can create duplicates?
- An HTTP status code starting with 5, like 500, means what?
- In a relational database, what two rules does a primary key enforce?
- Database indexes speed up searches. What is their cost?
- A bank transfer takes money from Alice and gives it to Bob. Why wrap it in a transaction?
- In ACID, what does Durability promise?
- What is a common difference between NoSQL and relational databases?
- SQL is called a declarative language. What does that mean?
- In functional programming, what makes a function purely functional?
- What is the difference between a unit test and an integration test?
- According to MIT's software course, what should you do first when hunting a bug?
- What caused the Y2K problem?
- What happens to a signed 32-bit Unix time counter in January 2038?
- On a website, where does backend (server-side) code run?
- After DNS finds the server's IP address, how does the browser open a TCP connection?
- While building a page, what makes the browser pause parsing the HTML?
- What does the HTTP header Cache-Control: no-cache actually mean?
- Your browser asks if its cached file changed and gets 304 Not Modified. What happens?
- Which three goals form the CIA triad?
- Which password property most directly increases resistance to guessing?
- What does current NIST guidance say about password composition rules?
- What should a service check when a user creates a password?
- Why is a unique salt added before hashing a password?
- What does bcrypt's work factor control?
- What makes authentication truly multi-factor?
- Why can a WebAuthn passkey resist a lookalike site's login prompt?
- What separates phishing from an ordinary unwanted message?
- How does a classic computer virus replicate?
- What distinguishes a classic worm from a virus?
- What defines a Trojan horse program?
- What does file-encrypting ransomware do before demanding payment?
- What does the principle of least privilege require?
- Which step makes patch management more than simply installing updates?
- What makes an attack a zero-day attack?
- What key does symmetric encryption use to encrypt and decrypt?
- What keys make asymmetric cryptography asymmetric?
- Which keys create and verify a digital signature?
- What does a TLS 1.3 key exchange establish for protecting data?
- Who can decrypt an end-to-end encrypted MLS application message?
- What can a messaging provider sometimes learn despite encrypted content?
- Where does a remote-access VPN tunnel's protection end?
- What happens to ordinary internet traffic during split tunneling?
- Which information commonly drives a basic firewall ruleset?
- What extra context does a stateful firewall remember?
- How do parameterized queries block SQL injection?
- Why must XSS output encoding match the browser context?
- Why should at least one backup be isolated from production systems?
- What does a restore test prove that a backup job report cannot?
- Why is a mirrored RAID array not a backup?
- About how much downtime does 99.9% monthly availability allow?
- What reliability problem is redundancy designed to remove?
- What must happen during a successful failover?
- Why should every logging host use a common time source?
- What does a SIEM add beyond collecting log files?
- Which activities belong in an incident-handling capability?
- What is the immediate goal of incident containment?
- What does incident eradication remove after containment?
- How can a service verify an age threshold while collecting less data?