David Hemphill is Chasing Dopamine

Matt introduces David Hemphill as a Laravel community OG, co-creator of Laravel Nova, creator of the Beard CSS framework, and someone connected to Tailwind’s origin story. He also describes David as a fighter, keyboard enthusiast, guitarist, singer, and songwriter. David explains that his fighting background is in mixed martial arts, including striking and wrestling with small four- or eight-ounce gloves rather than boxing gloves. The gloves primarily prevent severe cuts; they do not provide much hand protection. David entered MMA roughly two decades ago after hearing from an associate pastor that a gym connected to Pat Miletich, an important figure in the UFC’s early history, was opening in his small hometown. The decision seemed out of character: David had not played sports, was extremely thin at about 125 pounds at six feet tall, and had been voted “most artistic” in high school. He had mainly played in bands. Still, he was drawn to the adrenaline and dopamine rush. Early training felt like secret knowledge, especially when he submitted a roughly 6-foot-5, 230-pound man while weighing only 125 pounds. David says the submission forced the larger man to choose between having his arm broken and tapping out, which convinced him that the techniques genuinely worked.

After about a year of training, David began accepting opportunities to compete in small, often unsanctioned shows. These events could take place in bars or in another gym’s territory, with no athletic commission overseeing them. His first win was especially addictive: his opponent failed to answer the bell for the second round because he could barely see through one eye. David says fighters primarily care about their win-loss record, not elaborate statistics. Depending on the level, bouts might use three three-minute rounds, three five-minute rounds, or five five-minute rounds; judges generally assess aggression and who held the more dominant positions. David’s first loss came in Clinton, Missouri, against the local high-school wrestling coach, in the coach’s own gym. David repeatedly put the opponent in an arm bar and heard his elbow pop, but the wrestler refused to tap. David ultimately lost by decision, even though the referee—an active UFC fighter—told him afterward that he thought David had won but had spent too much time on the ground. David then lost another bout by legitimate submission and briefly questioned whether he was actually good enough. He eventually built a 9–3 record, with his last fight in 2019 at age 37. He says age, recovery time, and training against younger, hungry fighters eventually made continuing less appealing. Fighting gave him aggression release and adrenaline, but he valued it mainly as a physical and mental chess match: creating openings, controlling positions, and solving danger without wanting to hurt anyone.

Matt observes that David has maintained several interests—fighting, technology, keyboards, and music—despite both of them having ADHD and experiencing periods of intense focus. David says the interests do sometimes churn, but music was his first real love. He started guitar at 15 or 16, played in church, and formed mostly punk-rock bands influenced by Green Day, Blink-182, and Rancid. His bands later moved toward a broader rock sound and tried to book shows. The appeal was similar to MMA: everything could go wrong, but when the performance locked in, the result produced an intense rush and a powerful focus state. David recalls having a band that played major festivals and opened for bands they admired, though he wondered whether they were truly good or simply the best local choice for those events. The band eventually dissolved after he moved across the country. Matt relates this to his own experience as a bass player, watching former peers go on to professional careers after he stopped playing more than a decade ago. David still writes, records, sings, and plays guitar, bass, drums, and some keyboard. He can build a complete production himself, including programmed drums, and sees music-making as another way to pursue focused, high-pressure creative work.

David describes using Suno to test musical ideas before committing to a full production. He might begin with a 20-second melody recorded in a phone memo or a demo consisting of acoustic guitar and vocals. He can then upload the idea, identify sections such as verses, choruses, and instrumentals, and specify styles, vocal type, and how creatively the model should interpret the material. He may generate roughly 20 variations, changing the prompts and parameters until he finds an arrangement that suggests the song is worth pursuing. David generally does not release the Suno output unchanged. Instead, he downloads a useful version, brings it into a digital audio workstation, decomposes it into drums, bass, keys, and guitars, and records his own versions of those parts. He compares the process to building an airplane while it is flying and recognizes the unresolved question of when a song becomes “my song” rather than an AI song. He views AI as another tool, much as programmed drums and sampled instruments became accepted tools. However, he does not want a diet of entirely AI-generated music. He argues that listeners still want humanity in the work, and that the challenge is preserving that humanity inside an increasingly automated process. He and Matt also note that AI-generated background music may be harder to distinguish when listeners are not paying close attention, but repetitive, generic music becomes obvious when its sameness is noticed.

David says his use of AI for programming resembles his music workflow: it lets him act on an idea quickly and determine whether it is worth pursuing. At Zillow, where he works, AI-assisted code generation is encouraged. He has been generating code with AI for almost two years, which he acknowledges is aggressive compared with developers who are only beginning to relinquish direct authorship of every line. His role, in his description, is no longer primarily that of a code writer. Code is an artifact of his thinking and prompting; the more important work is deciding what to build and ensuring that it is performant, testable, secure, and correct. David still believes engineers need fundamentals such as concurrency, race conditions, and security, even if they no longer need to remember every syntax detail or the argument order for an array function. He initially resisted AI because he saw people operating “lights-off software factories,” submitting requests and shipping generated code without inspecting it. He compares hand-writing code to farming with hand tools: satisfying and appropriate for a personal garden, but not scalable for industrial production. Matt adds that this expectation is not universal—many companies still prohibit sending code to cloud AI tools—but agrees that fast-moving organizations increasingly expect developers to use AI and deliver more quickly. David sees mass-produced code as an industrial revolution for software, with boutique, human-written code remaining a niche offering.

David’s mechanical-keyboard interest follows the same pattern as his other hobbies: he enjoys discovering a system, refining it, and chasing the focus and dopamine it provides. He describes the old group-buy model, in which an individual organizer would design a keyboard, publish renders and specifications, collect orders, and coordinate production with factories. Buyers might pay around $300, choose layouts and features, and then wait two or three years while cases were milled or injection-molded, PCBs were produced, and the parts were assembled. David estimates that he participated in about 15 group buys and now has close to 30 keyboards. Some projects were delayed or failed entirely; organizers sometimes ran out of money, went bankrupt, or disappeared. The resulting keyboards may arrive as a case, PCB, and switch plate rather than a ready-to-use device. David chooses switches, keycaps, stabilizers, dampening materials, and other components, and has sometimes soldered switches into place. Some designs use a separate daughterboard to move the USB connection for ergonomic or layout reasons. He enjoys comparing linear, tactile, and clicky switches and describes the process as similar to developing preferences for wine or bourbon. His “endgame” is roughly a 65-percent layout, though his preferred keyboard changes with the day. He may use a custom board, an off-the-shelf Omnitype Bauer Lite, or even an old Model M with a USB adapter. For a newcomer, he recommends a Keychron 75-percent keyboard with function keys and navigation keys; Matt recommends the Keychron K2 and K3 Max for general and travel use.

David strongly praises his Tesla while separating that enthusiasm from Matt’s explicit dislike of Elon Musk. The primary benefit, he says, is that the car can drive itself for most of long trips. On drives to Dallas, he would rather maintain a small amount of attention while the car handles much of the road than manually drive for hours while becoming exhausted. He also describes using the system during a chaotic family trip to a pumpkin patch, when he could monitor the road while handling food and other distractions. He believes the cars are safer and could eventually be useful for his children, especially because speed and acceleration limits can be configured. He also enjoys Tesla’s playful features, including fart sounds, Halloween graphics, zombies, and a skeleton replacing the normal driver representation. David’s newest interest is fishing, which combines childhood memories with another opportunity for concentrated attention. He fishes in a neighborhood pond that predates the surrounding subdivision and appears to be maintained with aeration and fountains. Although he expected small sunfish, he has caught large bass and enjoys not knowing what will take the lure. He has become interested in rods, reels, gear ratios, lure types, and rod action, and uses the Fishbrain social network to record catches while protecting his best fishing spots. Fishing offers a quieter version of the same reward he finds in fighting, music, programming, and keyboards: sensitivity to small changes, focused attention, and the dopamine hit of discovering that something is happening on the line.