March 1987
March came with a thaw that felt like permission. Snow receded into dirty gutters, and New York's mood shifted from hunkered-down survival to forward motion. In the Workshop, the thaw registered as a loosening of deadlines—enough room to breathe, enough daylight to plan bolder moves. Julian arrived at his desk early, ledger open, a cup of coffee cooling at his elbow. The chart on the mezzanine had new pins: recent minority acquisitions in regional distribution, the Queens tournament report, a small box labeled European Outreach where Marcus had folded the French distributor proposal. He skimmed the numbers, then closed the book. Numbers were necessary; structures were necessary. But language — the way people, machines, and cultures translated power — was becoming the real frontier.
He had been carrying the idea for weeks, half-formed, half-remembered: a computing language rooted in Sanskrit, engineered to be compact, logically consistent, and expressive enough to serve not only software but protocols for secure, efficient communications across hardware architectures. The frozen archive — the mind-internet he accessed by thought — had brought surfaces of articles, forum threads, and academic papers from his previous life that suggested natural languages with precise grammars made for computation. He had not spoken of it. Not yet. Now, with a holding company, a distribution network, and pockets large enough to absorb early research costs, the time felt right.
He walked down to Anna's lab while the city still hummed with early commuter traffic. The lab smelled of solder and solvent, and a whiteboard held equations that looked more like poetry than electronics. Anna looked up from a bank of oscilloscopes when he entered. She didn't ask him to sit; she already knew why he was there.
"I want to build a language," he said simply. "A computing language built from Sanskrit's logic. Compact, efficient, and readable — not only by machines but by human scholars. It will be used in embedded systems: phones, routers, and eventually satellites. Not for vanity. For utility."
Anna sat back, weighing him with eyes that had learned to see through strategy into feasibility. "We still can't mass-produce stable boards. You want to build a new language?"
He smiled at her exactness. "Two projects in parallel. We keep the material camouflaged as stage equipment. But we start an R&D group: linguists, compilers, cryptographers, and software engineers. We prototype the language and implement a small virtual machine that runs on our current test boards. The VM will be minimal — code size tiny, deterministic. If we pull that off, the rest is engineering."
Anna tapped her pen. "Where do you find Sanskrit programmers?"
Julian's eyes rolled toward the mezzanine map where India had its own cluster: Raga research hub, Lotus cultural centers, a dotted line toward a planned campus in Bengaluru. "We have partnerships in place in India. Marcus has been quietly cultivating relationships with a few academics in Delhi and Madras. Bring them in; pay them market rates plus incentives tied to implementation milestones. Hire local graduates. Make it meritocratic. No politics."
She exhaled, a laugh that was both weary and exhilarated. "You ask the lab to host linguists and computer scientists, and you give them stage rigs to hide behind."
"Camouflage," Julian corrected. "And redundancy."
Within a week, Anna converted a corner of the lab into a software room: stacked terminals, a whiteboard for grammars, and a locked cabinet of test boards. Julian instructed Sophia to draft NDAs, fund transfers, and consultancy agreements in a way that kept the project legally sheltered under Lotus Lighting Systems. He insisted the documentation match the theatrical cover story: purchase orders for lighting designers, invoices for theater fixtures. Sophia objected only once, on principle. "We're inventing a legal sculpture," she said. "Make sure the sculpture doesn't crack." Julian promised she would be the sculptor.
The first hires were two young linguists from Delhi: Dr. Anjali Rao, who had done work in formal grammar theory, and Professor Vikram Sen, a calming, precise man with decades of teaching in Madras. Along with them came a former compiler engineer from a small Cambridge startup, a cryptographer who had written quietly influential papers on compact key exchange, and three junior software engineers who could code long into the night without noticing the hours. Julian met them in the Workshop one evening, watching them cluster around Anna's whiteboard as if the equations were a different language they were eager to unpack. He didn't lecture. He listened, asked a few questions, and then let them find the music in the syntax.
He used his mind-internet sparingly but purposefully. When Anjali raised a question about a morphology edge case that might bloat the parser, Julian closed his eyes and let the frozen archive surface: an obscure 1982 paper on minimal grammars for constrained processors, a forum thread about stack memory in early microcontrollers, a snippet from a lecture on finite-state machines. He quoted the insights without context, and the team turned them into experiments. The mind-internet gave him patterns; the team turned patterns into prototypes.
Weeks became a rhythm. Morning sessions were for grammar workshops: designing primitives, the equivalent of nouns and verbs for machines—operators, state transitions, and a minimal type system that mirrored Sanskrit's clarity. Afternoon sessions were compiler work: lexical analysis, code emission, and register allocation tuned to the idiosyncrasies of their test boards. Night sessions were for cryptography: embedding secure key exchange primitives into the VM without making compilation complex. Julian wandered through these sessions like a careful gardener, pruning when needed, letting the shoots find their light.
Outside the lab, the rest of the empire marched on. Lotus continued to distribute films; the indie strategy paid off in accrued credibility. A small distributor in Ohio had agreed to a pilot with Vanderford Distribution; attendance reports trickled in and showed interesting patterns—elevated repeat attendance in neighborhoods where Lotus hosted community Q&As. Marcus forwarded graphs of retention curves and ROI estimates. He had also been stingy with interviews; he preferred results to rhetoric. At the Workshop's weekly review Julian took the numbers, nodded, and then shifted the conversation back to the language project. He wanted cross-communication between industries: seed programmers from Raga to write libraries for media encoding; integrate the language into Lotus's editing tools as a scripting language that could one day replace machine-specific plugins. It was about building ecosystems, not solitary products.
Julian's strategy included a careful financial plan. He authorized an initial R&D fund of $4.5 million, drawn from the trust's reserves and a small line of short-term credit that Marcus negotiated at favorable terms with a New York private bank. The numbers were tall and risky; Sophia ensured that legal protections, escrow accounts, and staged funding releases guarded against leaks and governance challenges. Julian insisted on keeping the company private—no IPO, no public float. Ownership remained concentrated: the trust held 75 percent; 20 percent was allocated to key employees under clear vesting schedules, and 5 percent reserved for strategic partnerships. Investors would come later, if at all; his priority was control. Control, he thought, was less about ownership and more about direction.
As the language prototype took shape — a grammar dubbed Sāmrta in early notes, a nod to the Sanskrit sāmarthya meaning capability — they faced the first real technical trial. Sāmrta's interpreter needed to run within a tight memory budget and provide atomic operations for secure messaging. The prototype initially required 64K of memory to run comfortably; their target boards offered half that. Anna's team experimented with micro-optimizations: tail-call elimination, compressed instruction encodings, and a microkernel that offloaded certain responsibilities to hardware interrupts. The frozen archive was useful again: Julian recalled clever register allocation tricks from an old embedded systems paper. They adapted the tricks; the interpreter shrank.
They called the first successful run a quiet victory. A small test message — a signed, encrypted token that would later serve as a handshake between a phone and a network node — was created, serialized in Sāmrta bytecode, and executed on the test board. The message held, decrypted and verified. Julian felt the small, bright stab of triumph he had learned to savor: not a popular headline, but a precise, practical win. "We have a language between machines," Anna said softly, and for once her voice carried the awe Julian had kept for private ledger entries.
He recorded it in his ledger the way he recorded all things: short, deliberate lines. Language is leverage. A language that fits machines and minds bends reality.
That month also deepened the diplomatic work. The French distributor's pilot showed promise; later, a Belgian audiovisual cooperative asked to license a batch of regional documentaries. Julian negotiated with careful hands — IP protection clauses, nondisclosure on the technologies behind potential mobile devices, and revenue splits that favored reinvestment into R&D. While the world saw Lotus as a small player making clever indie moves, Julian planted deeper roots: distribution channels that would one day carry not only films but devices, software, and cultural exports.
Yet for all the engineering and negotiation there were cultural stitches Julian refused to ignore. Raga organized a small music residency that month, bringing traditional Indian musicians to New York concert halls. The residency doubled as a testing ground for Sāmrta in media metadata: encoding ragas, talas, and artist lineages in machine-readable, culturally respectful formats. Anjali worked with Raga's cultural strategists to ensure the language preserved nuance: a note's ornamentation, the microtiming of a tala — encoded not as crude numbers but as expressive primitives. It was research and cultural diplomacy at once, and it pleased Julian that the empire's technological ambitions entwined with cultural preservation.
The personal strain, meanwhile, crept into small places. Demi's name began to appear in better columns; the Providence festival exposure led to an offer for a supporting role in a larger independent film. Her schedule tightened, and late nights at the Workshop became less frequent. When she was in town, they shared dinners where scripts and ledgers intertwined in conversation. When she was on set, Julian sent notes: a line read, a pacing suggestion; he never suggested edits that compromised the director's vision, but he used his taste to advise. Their relationship was not a headline; it was a pragmatic companionship that allowed both to move forward without being consumed.
By month's end Sāmrta had an alpha release: a minimal syntax, a bytecode emitter, and three small libraries — a serialization module, a crypto handshake, and a compact string type tailored for Indic scripts. The team had ported a tiny interpreter to a vendor's test board, and the run in the lab lasted a week under continuous stress tests. It wasn't production-ready. It was fragile. But it worked.
At the Workshop's final review for March, Julian asked each leader to place one small, concrete objective on the board for April: Marcus would finalize deals with two more regional chains; Anna would harden the interpreter against memory leaks; Anjali would prepare a whitepaper showing Sāmrta's formal grammar and a set of use cases for embedded devices. Sophia would finish the legal scaffolding for R&D funding. Mira would plan a cultural showcase that married Lotus's films and Raga's music in a weekend festival that would double as a soft-launch for metadata experiments.
They worked late into the evening, lights in the Workshop like steady beacons. When the team finally dispersed, Julian stayed to write. In the ledger he added a new line, short as ever, but with the weight of months: Build language. Build trust. Make India's voice a protocol, not a slogan.
Outside, the thaw had become a steady trickle. The city smelled faintly of wet asphalt and opportunity. Julian closed the ledger, turned the page, and allowed himself a moment of cautious satisfaction. The Workshop was not merely building companies; it was building the grammar for a future where culture and code spoke the same truth.