#Commercial Workbench | Explore Tumblr posts and blogs

mariacallous · 7 months ago

Text

On the top floor of a building somewhere in Ukraine is a drone workshop.

Inside is a chaotic workbench covered in logic boards, antennas, batteries, augmented reality headsets, and rotor blades. On one end of the room is a makeshift photo studio—a jet-black quadcopter drone sits on a long white sheet, waiting for its close-up.

This particular workshop’s Geppetto is Yvan. He grins as he shows off his creations, flittering around with a lit cigarette in his mouth, dangling ash, grabbing different models. (Yvan is a pseudonym; WIRED granted some of the people in this story anonymity due to the security risk.)

Yvan holds up a mid-size drone: This model successfully hit a target from 11 kilometers away, he says, but it should be capable of traveling at least 20. He’s trying different batteries and controllers to try to extend the range. He screws on a stabilizer tailpiece to a hard plastic shell—Yvan 3D-prints these himself—and holds up the assembled bomb. It’s capable of carrying a 3.5-kilogram explosive payload, enough to take out a Russian tank.

He uses his index finger and thumb to pick up a nondescript beige chip: This, he says, is what he’s really proud of.

One big problem with these drones—which are based on commercially available first-person-view (FPV) or photography drones—is that their explosive payload is jimmy-rigged on. It requires the drone to crash in order to close the circuit and trigger the explosion.

This chip, Yvan says, allows for remote detonation from a significant distance, meaning the operator can park their drone and lay in wait for hours, even days, before it goes off. He expects this technology could, eventually, be connected to AI—exploding only if it registers a nearby tank, for example. He has created a long-range smart land mine, I note. After the idea is passed through our translator, he nods enthusiastically.

There are many of these FPV drone workshops around Ukraine—Kyiv estimates there are about 200 Ukrainian companies producing aerial drones, with others producing land- and sea-based uncrewed vehicles. But Yvan, grinning proudly, insists that the manufacturer which he represents, VERBA, is the best.

Ukraine is facing increasingly tough odds in its defensive war against a better-resourced, better-equipped enemy. Thanks to delayed aid from Washington and shortages in other NATO warehouses, Ukraine has lacked artillery shells, long-range missiles, and even air defense munitions.

These drones, however, represent a bright spot for the Ukrainians. Entrepreneurship and innovation is scaling up a sizable drone industry in the country, and it’s making new technological leaps that would make the Pentagon envious.

The age of drone warfare is here, and Ukraine wants to be a superpower.

After Yvan showed off his workshop, we loaded into the car to visit one of his factories.

Behind a steel door is a room filled with racks, where 30 3D printers are working simultaneously, printing various drone components in unison. The twentysomething employees seem accustomed to the screeching alarm—some are soldering the drones together, others are tinkering with designs in AutoCAD, one is lounging on a sofa.

Strung across one shelf of 3D printers is a black flag, a take on Blackbeard’s (apocryphal) pirate flag. It shows a horned skeleton wearing an AR headset and holding a controller, thrusting his spear toward a bleeding heart as a quadcopter flies above.

In the first year of the war, when FPV drones were providing extraordinary footage of the front lines and viral video of unmanned aerial vehicles (UAVs) dropping grenades on Russian tanks captivated the world, Ukraine was snatching up every consumer drone it could find. Chinese technology giant DJI became a household name in Ukraine, thanks to its drones’ ubiquity on the front lines. Ukraine’s early advantage was quickly lost, however, as Russia scrambled to snatch up these Chinese-made UAVs.

“When Russia sees, from Instagram, my product, Russia starts buying all these components in China,” a VERBA executive says. The new demand from Moscow can often cause either shortages or inflation, squeezing out the Ukrainian companies. So entrepreneurs like Yvan began building their own.

When Yvan began his operation in the early months of the war following Russia’s February 2022 full-scale invasion, he was creating a handful of frankendrones to send to the Ukrainian Armed Forces. Now, Yvan says, his operation is producing 5,000 FPV drones per month. He offers a range of systems, from a mammoth 12-inch model to a 4-inch prototype.

At first, these entrepreneurs were pursuing this project on their own—scrambling, like most of the country, to be useful in helping Ukraine defend itself. Kyiv was initially cool to the idea that a domestic drone industry was worth the money and attention, especially given the demand for more conventional arms. Some in the military, one executive says, dismissed the utility of these innovative weapons and surveillance platforms as merely “wedding photography drones.” (One executive said Oleksandr Syrskyi, Ukraine’s new commander in chief, had been an early adopter inside the military, directly contracting 10 firms in early 2023 to begin assembling new technology for his forces.)

That attitude changed in 2023, when Ukraine set up Brave1, a government-run technology agency and incubator that helps connect private enterprise to the Ukrainian Armed Forces.

Since its creation, Brave1 has worked to streamline design, development, and procurement of new defense technology, while helping companies navigate government and military bureaucracy. Brave1 has already awarded more than $3 million in research and development grants and connected more than 750 companies to the Ukrainian Armed Forces.

When United24, the Ukrainian government’s in-house crowdfunding platform, first pitched an “army of drones” to its donors in 2022, it aimed to buy just 200 units. Today, Ukrainian president Volodymyr Zelensky projected late last year that his country would produce over 1 million drones in 2024.

“I would say that we can even double this number,” Natalia Kushnerska, head of Brave1’s defense technology cluster, tells WIRED.

“We have the responsibility and the motivation to do it today and to do it very fast,” she says. “Because we don't have any other choice.”

This is a war, one executive told me, “where the economy matters.”

Even hampered by sanctions, Russia has a $2 trillion economy—about 6 percent of that is geared toward its wartime production. Ukraine’s entire GDP, by contrast, is less than $200 billion.

While Kyiv has received substantial support from its NATO partners, it faces constant pressure to find efficiencies. The economics of these drones are looking better and better.

Yvan’s drones are, compared to conventional munitions, cheap. His most expensive unit runs about $2,500, but the cheapest is only $400.

Early in the war, the Ukrainians could reasonably expect—depending on weather, the mission, and Russian jamming efforts—that about 30 percent of their drones would connect with the target. Today, good Ukrainian-made systems are approaching a 70 percent success rate.

It can often take four or five artillery shells to successfully destroy a medium-range target, such as a tank. At $8,000 per shell—which are in short supply and high demand—that is an expensive proposition. Even if it takes two of Yvan’s most expensive drones to achieve the same objective, that’s thousands of dollars in savings. The proliferation of these drones reduces the “cost-per-kill,” as one executive phrased it, and reduces the strain on those dwindling ammunition stockpiles.

Even if Yvan and other producers are making more and more of their systems in Ukraine, they still rely on Chinese suppliers for critical onboard components. That comes with a trade-off—Chinese suppliers are cheaper, but they tend to be of lower quality and are happy to do business with Russia as well. Other options, such as companies in Taiwan, the United States, Canada, or Europe, are better quality but can be several times more expensive.

These supply chains, Yvan says, are “complicated.” Drone manufacturers who spoke to WIRED say anywhere between 40 percent and 80 percent of their drone components are made in Ukraine. Asked how long it would take before Ukraine manufactures nearly everything in these drones, from the rotor blades to the onboard components, Yvan provides a bullish estimate: “six months.”

It’s not an entirely unrealistic dream. Mykhailo Fedorov, Ukraine’s deputy prime minister and also minister responsible for digital transformation, said late last year that Kyiv hopes to break ground on a semiconductor factory, capable of producing 50,000 chips a year, by 2025. Ukraine produces about half the world’s supply of neon, necessary for the lasers used to make the chips.

There are already companies in Ukraine that have developed electronic design automation software—a necessary tool for producing chips—and that do electronic assembly inside the country itself. An industry source tells WIRED that a working group was formed in late 2023 to chart out how Ukraine could be a player in the semiconductor industry.

Another defense technology executive, Igor, manufactures considerably more-sensitive drones. “We definitely don’t buy anything from China,” he says. His products are more expensive, he says, “but we are looking for something that would differentiate us from the Russians.” At the moment, he says, “Russia is ahead.” He’s hoping to close that gap.

For any of this to work, however, there needs to be demand for these drones. The more they can sell, the more they can invest. “The things that they need,” Kushnerska says: “contracts and money.” Demand has certainly grown—fundraising platform United24 helped finance a fleet of naval drones and raised funds to purchase 5,000 surveillance UAVs. Other organizations have led similar purchases. The drone-makers, however, say it’s just not enough.

In early 2023, Ukraine’s parliament passed new laws to regulate how drone manufacturers can contract with the state; while profiteering is generally discouraged in the wartime economy, the law specifically allows the companies to charge up to 25 percent profit.

Yvan says he charges just a 10 percent premium for his drones and reinvests all that profit back into his operation. Representatives from other drone companies who spoke to WIRED say they operate on a similar basis.

More orders will mean more investment. Thus far, NATO countries have preferred to purchase locally-made equipment and ship it to Ukraine. That may be changing.

Bill Blair, Canada’s minister of defense, visited Kyiv shortly before I was there. While there, he announced that Ottawa would donate 800 Canadian-made drones to Ukraine. While the donation was lauded, a senior official asked the minister, “Why didn't you buy our drones?” After being briefed on the various innovations taking place in the Ukrainian drone industry, Blair was convinced. “We're also going to find ways to invest in Ukrainian industry,” he tells WIRED. “The point of the [Ukraine Defense Contact Group drone coalition] is to create capability, not only in the countries that are in the coalition but also capability in Ukraine.”

Even still, bureaucracy moves slowly. What’s more, startups—some of which are helmed by technologists or special effects gurus with no experience in procurement, let alone war—are often learning as they go. One executive, covering his eyes with his hand, says: “It’s like going completely blind.”

Not every company has been able to hack it. One executive says he’s aware of five defense technology startups that have shut down since the war began.

Much attention has been paid to FPV drones. They reinforce the idea that Ukraine’s defense is a scrappy, homespun effort. But even as the country has professionalized production of these light, agile drones, it has rapidly spun up production of other, more complicated systems.

One of Ukraine’s biggest disadvantages, from the start of the war, has been its difficulty in hitting targets inside Russia. Because Moscow has so effectively dominated the skies, Ukraine has been left playing defense.

That equation has changed substantially in recent weeks. Ukraine has had enormous success in attacking Russian oil refineries—knocking out as much as 15 percent of the country’s total refining capacity—and bombing Russian air bases. This has all been made possible by Ukrainian-made long-range attack drones.

Igor, who represents a company responsible for producing those long-range bombers, says they have developed a unit capable of flying 1,000 kilometers and carrying a 25-kilogram payload and has produced “several hundred” units for the Ukrainian Armed Forces. And they are actively working on a new model, capable of flying up to 2,500 kilometers. (It will pack a smaller punch, he said: “The longer you go, the lighter the payload.”)

These systems are more expensive: from $35,000 to $100,000. But if they can destroy millions of dollars worth of Russian equipment, that’s a bargain.

“These are no simple drones,” Igor says. “We don’t have the luxury, like the Western guys, to spend years in development.”

They’re not stopping with drones, either. They’re using the same technology to develop Ukrainian-made missiles, capable of flying farther and doing more damage to Russian military infrastructure, tucked well behind the front lines, which is regularly used to attack Ukrainian cities.

Igor’s goal is to “bring the war to Russia.” FPV drones have broadcast the realities of the front lines in high definition—long-range bombers could successfully make it feel real, he says. “They don’t suffer like we suffer.”

The effort to bring the war to Russia is advancing on multiple fronts. One of the most famous uncrewed systems of the war has been Kyiv’s Sea Baby drones. Videos have gone viral of these sleek ships clipping along the waters of the Black Sea.

According to Kyiv, they can carry 850 kilograms of explosives, go 90 kilometers per hour, travel some 1,000 kilometers—and they are invisible to radar. This is the kind of capability that the Pentagon, and other defense departments, has spent years trying to develop. “We like to joke that everything we do now, in Ukraine, takes three days—globally, it takes three years,” Brave1’s Kushnerska says.

Ask around Kyiv about these drones, however, and everyone is mum. Even otherwise talkative defense sources go quiet when asked about the Sea Babys. Asked about the vehicles, one defense executive smiled and said simply, “That’s classified.”

Kushnerska is equally evasive: “We need to keep silent about new solutions and new surprises that we are preparing for the enemy.”

The skullduggery is understandable. These uncrewed vehicles have been responsible for doing massive damage to Russia’s prized Black Sea fleet and spearheading the first major attack on the Kerch Bridge, in Crimea, in 2022.

Developing naval drones, however, is relatively easy compared to uncrewed land systems.

Over tea with Stepan, another defense entrepreneur, he lists the litany of difficulties of trying to build uncrewed land systems: They don’t travel well over tough terrain, they don’t operate well in inclement weather, and they don’t tend to go very far.

And yet, Stepan says, his company has overcome all those obstacles—which the Pentagon is still wrestling with—and has put these land systems in the field. Plus, Stepan says he’s “pleasantly surprised by how they’re being used.” He says their smallest unit, which has generally been used to deliver food and equipment, recently rescued and evacuated a wounded soldier from the front line.

Ukraine is not the only side deploying these land systems, however. In late March, pro-Kremlin channels celebrated what they said was the successful deployment of Russian-made uncrewed land systems, outfitted with an AGS-17 grenade launcher.

Ukraine believes its advantage will come from how it dispatches these systems. “You need a mesh system,” Stepan says. And that’s one of the single hardest things to do. Ukraine has started dispatching repeater UAVs, which are used to extend the base station signal, allowing the drones to fly farther and defend better against Russian jamming.

One ground drone, basically a mobile machine-gun turret, boasts an 800-meter range. What’s more impressive, however, is what happens when the land system is paired with a surveillance drone. Rather than them firing directly ahead, Stepan’s team has been training Ukrainian soldiers how to raise the weapon's trajectory, firing in a parabolic pattern and using the drone’s camera to adjust its aim. This tactic, he says, extends the drone’s firing range to 2.4 kilometers.

Doing combined operations with a couple of drones is hard enough. If Ukraine wants to really take advantage of these autonomous systems, it will need to figure out how to command multiple systems across land and air—and that’s where artificial intelligence comes in.

Stepan walks through the four levels of how AI can augment warfare: One is reconnaissance, where machine learning can be used to collate large volumes of footage and satellite imagery. Two is “copiloting,” as he calls it, where AI can analyze that intelligence and help draw insights. Third is planning, where AI can help develop “interlinked, complex orders” for multiple systems across land and air; he likens that to having AI develop football plays. Finally, step four is full autonomy, where AI collects intelligence, analyzes it, develops orders based on the intelligence, and dispatches and commands autonomous units based on that information—although humans review and approve each step of the process.

There are steps beyond this, Stepan notes, that remove human involvement entirely, but he isn’t interested in going there. Another executive recounted a story of how one company designed an autonomous machine gun, capable of conducting object detection and opening fire on its own—that was a “big, big problem,” he says, after the weapon’s radio signals were jammed and it began firing wildly. “I think we can do this slowly,” he adds.

Stepan’s systems are capable of operating at step four, he says. It means his systems have the “ability to take in variables” in real time—it allows his drones to change tactics depending on the environment. He provides examples: “What if our team is close? What if there is [electronic warfare]? What if one system loses connection?”

Kushnerska says Ukraine, alive to the concerns about and risks of AI on the battlefield, is mostly interested in using artificial intelligence only in the “last mile.”

It’s not enough to build drones. Ukrainians also have to know how to pilot them.

The last stop on Yvan’s tour is at a strip mall some distance away. Outside, a group of fresh-faced young men smoke cigarettes and enthusiastically greet him as he walks past.

Inside is a sterile classroom, with a dozen desks laid out—each featuring a tablet, a workstation, and an array of tools. In the back corner are pallets of FPV drones waiting to be unloaded.

This is Yvan’s drone school. Here, students learn not just the ins and outs of piloting these quadcopters but also how the machines work and how to repair them. Down the hallway is a large conference room where the students first test their skills—flags and checkpoints are propped up on cardboard boxes taped together into platforms of different levels. Once students can successfully navigate this makeshift course, they graduate to piloting the drones outside.

Yvan’s drones are normally painted jet black, designed to look as nondescript as possible. One drone, sitting on a desk in the training school, is spray-painted a bright orange. Yvan grins: “We’re sick of losing them in the grass.”

As Kyiv mobilized tens of thousands of ordinary Ukrainian men to fight, training has been a critical necessity. Particularly as ammunition supplies have dwindled, virtual training has been especially attractive. High-tech combat simulators have allowed Ukrainian troops to simulate real combat scenarios with rifles, rocket-propelled grenades, even anti-tank missiles. Ukrainian entrepreneurs are hoping to have dozens of these simulators online in the near future, with the goal of training 100,000 troops.

An industry source tells WIRED that a drone combat simulator went online last month, allowing trainees to simulate the entire process of launching a long-range drone strike. Version 2.0 is being rolled out now, they say, adding that it is likely the first immersive offensive drone simulator in operation. The simulator is also intended to help Ukrainian pilots practice integrating their drones with land systems, which is notoriously difficult for even experienced soldiers.

While Yvan’s drone school offers hands-on experience for users of the FPV drones, this new drone simulator allows pilots to practice long-range targeting, flying in adverse weather conditions, and countering electronic warfare.

All of this—the FPV drones, the long-range bombers, the flight simulators—is Ukrainian innovation at work. And it is moving remarkably fast. Some day, after the war is over, Yvan may well be on the front lines of a Ukrainian technology renaissance, fulfilling orders for the Pentagon. First, both he—and Ukraine—need to survive.

In recent weeks, Russian forces have made modest but steady advances along the front lines. Defense executives, meanwhile, see sabotage and industrial espionage as constant problems. Even more acute is the threat of Russian air strikes. One executive recently recounted how one of his company’s main facilities was nearly hit by two Russian cruise missiles. The risk is very real.

Leaving the school, Yvan opens up the back of his car. He rummages around and hands me two patches: One features a cartoonish and scantily clad woman wearing an FPV headset with the Ukrainian flag on the side, piloting one of Yvan’s rotocopters. The other, an army-green Canadian flag, carries the words “ALWAYS BE READY.”

5 notes · View notes

helloliriels · 2 years ago

Text

Sleepless in London (Pt.3)

“It was a million tiny little things, that…

Somehow, just meant we fit together?

And I knew ... the very first time we touched.

Like finding a missing puzzle piece, you didn't know you had lost ..."

John's voice drifted over the radio, barely having started to talk when the announcer cut in, and the host came back on:

“ … this is Dr. Irene Adler and we’ll be right back with John and Rosie, but for now, let’s open up the lines and take your calls …”

“Molly, is it absolutely necessary to have this drivel on?”

“It’s not drivel, Sherlock,” Molly defended, “It’s romantic!”

The pathologist was currently humming along with the ads playing on the radio, while dissecting a human brain; As Sherlock himself leaned in to check his samples more closely, from the workbench opposite. He closed his eyes in an attempt to focus, as the noise continued …

He had been careful to maintain his distance from the overly friendly and clearly unobservant resident of St. Bart’s for fear she would misinterpret any action on his part. And now she was playing a romantic talk show? Please.

“The romanticism on this show is entirely manufactured, Molly. While the highly exploitative nature of how she handles her callers is frankly, embarrassing,” he drolled, "if, by romantic - you meant to say humiliating and degrading ... then yes. This show is romantic.”

“So - what do you do for fun, then?” Molly asked, a cheery attempt to get him talking more … now that he had made the mistake of starting …

“Though we were having fun?” Sherlock replied, shutting her down.

Molly went back to her work, glum.

“Yeah, I’d just like to know how to get this man’s number?”

One woman was asking, right out the gate.

Sherlock snorted.

Molly glanced at him, and embarrassed, turned the program down a bit. The radio host was confirming that they would accept mail to the radio station, but would not give out any private data. She thanked the caller and accepted the next:

“John and Rosie, I just want to say, I'm here for you … !”

“Can we write to …?”

“How can I get in touch … ?”

“Umm, did he say how old he was? Like - I’m embarrassed to ask - but he sounds dreamy and I just - don’t wanna be disappointed?”

“Early-forties,” Sherlock supplied to the room at large, then with a snark added under his breath, “and far above your pay grade, clearly.”

The calls continued in this vein for a while … then they were re-playing the sensational bits of the program so far:

Rosie pleading,

“I think my daddy needs a new wife.”

John’s voice, responding,

“Why do you assume it has to be a woman?”

The Woman’s signature provocative style teasing,

“So we’re open to more choices then?”

Before cutting to another commercial break.

Sherlock could feel his eyes rolling back in his head, as he slammed his palm on the table. “MOLLY!” he begged. Pleaded.

(Pt. 3 long post, actually continued below cut)

Part 2 | Part 1 | Movie Cover | Subscribe on AO3

Molly Hooper lifted her scalpel and paused.

Meeting Sherlock’s glare with a kind, softened smile. “We’re the only ones here, Sherlock,” She reminded, returning to the task at hand with renewed zest, “the only two people in London bored enough, and lonely enough to be wasting our New Year’s Eve in a laboratory shared with dead bodies … ”

“I know,” Sherlock replied, grinning, “wish Christmas had been this lively!”

Her smile fell as she looked away, “Don’t you think … ? Maybe you could learn something from Dr. Irene?” She added with a frustrated huff, “about human feelings?”

“Human?” Sherlock blinked, poking at the human remains he was in charge of.

“Feelings?” She tried again.

They had already been there for four hours, and not only had Sherlock completely misunderstood her offer of coffee, he had also insulted her jumper, her lipstick and her hair care routine … On second thought, she reached over and turned the radio program back UP.

Sherlock rumbled,

“If you’re hoping that by listening to ‘Your Feelings and You’, I’ll suddenly catch emotions like the common cold … then I regret to inform you, your efforts are entirely wasted.”

“Wouldn’t be the first time,” Molly muttered under her breath.

But as she looked back at the gorgeous, curly haired detective - her breath caught still. He was prickly, and mean, and entirely too brilliant for his own good. But she had a feeling that two thirds of those traits were habits of self-defence …

Her hand fell on the small sprig of mistletoe tucked away in her pocket from last week, and sighed. She hadn’t been brave enough to get it out then, and it seemed pointless now … She chucked it in the bin and turned a flustered eye back to her work …

As the friendly doctor came back on the radio, and began to really converse about the woman he’d lost, and why he had loved her so very much …

It all sounded so romantic, and adventurous …

She even caught Sherlock listening in a few times, and crossed her fingers it was catching … despite his bluster and bluff.

“John … May I ask you one more thing?”

The radio host interpreted his silence as consent and continued, “how did you meet your wife?”

John had felt the tension fading, as he opened up to the radio host, her ease of conversing ... not knowing or really caring how long they had been on the air for. It felt so good to talk about Mary in this light … without all the weight of responsibility that usually came with it.

But with that personal reminder of their early memories ... it suddenly felt like she was prying.

He wasn’t sure if it had been the generous holiday spirit affecting him, before? Or perhaps the three or four pints he’d consumed this evening … trying to forget it all …

“Ooh, I don’t know if I can-!” John hedged, laughing. If only to shake the feeling off …

“Oh, you can -” Irene cooed, invitingly, “Please! I guarantee my listeners are hanging on your every word … “

“No, it’s … nothing like that -” John clarified, scratching at the back of his neck, “That is … I don’t mean to be rude-”

“-And I don’t want to pry!” Irene hastened to add.

“Yes you do.” John answered bluntly.

Sherlock felt himself responding out loud along with the grieving widower.

He paused.

His hand on the pipette, cold glass against the table as he listened more intently …

“It’s not a family friendly story …" John added, cautiously … looking down at his child, asleep in his arms ... “but seeing as Rosie is out for the count … and since everyone involved is now deceased …" he hummed, thinking on it further, "I'm not sure ... it may still be covered under the official secrets act …?”

“Are we talking … government secrets?” Miss Adler gasped, scintillating at the chance of higher ratings, “John? Were you and your wife secret agents, like Mr. & Mrs. Smith?”

“Oh, god, no!” John chuckled, “I’m afraid that part … that was all Mary …”

He chuckled thinking about how they really had met … that day in Türkiye …

The Ambassador's wife and her entourage, held captive in a bunker … John’s squad filing in, bandaging up the wounded and evacuating those that could be shifted to safer ground, after Mary had gone in and taken out the first real threats. How John had fired a shot and saved her life ... her meeting him bloodied and grinning to ... to receive treatment. Her hair the colour of spun gold.

“No.” He reiterated, “I’m no 007. Just a doctor. A good marksman back in my military days, but those are all over now, due to the-”

“I know this was not easy for you, John,” Irene cut in, "and we are running short on time. But i just wanted to thank you for being so receptive and open."

Sherlock felt his frustration growing at not being able to hear the other side of the phone line … To know what the man - the doctor - the military trained, excellent marksman - with a million more stories to tell than this woman could possibly fathom! Had wanted to say …

He wanted to climb through the phone and ask the rest of the questions himself … To know how John had been wounded and invalided out, ahead of his time ...? How he had ended up as a Doctor in London? And what he liked to do in his spare time … ?

It was an odd feeling … being curious about someone not tied to a case …

“When did your wife pass?” The Woman’s voice interrupted his thoughts again.

John’s reply took a moment …

“About a year and a half ago,” he said at last, getting very quiet.

Everything felt like it was tunnelling back in around him now … as John thought about the reality of the situation … alone, just as before … and his hand began to ache …

He flexed it and switched the phone to his other ear to relieve the pressure on his shoulder.

“Have you had any relationships? Since then?” She prodded.

“ No .”

The answer was blunt.

Rude. Intentionally.

Sherlock was hanging on every breath the doctor took … waiting …

Feeling the heat rising in the back of his own throat … even as John fought the lump rising in his own …

“Why not?” she asked them both (it seemed), softly.

“Irene … listen,” John began, “or should I call you Dr. Adler?”

“Dr. Irene, please,” The Woman suggested.

John and Sherlock both simultaneously rolled their eyes.

“We had a pretty tough time at first, I’ll admit that. But … we’re making do … I’m seeing a therapist - a real one -”

Sherlock smirked at the dig.

“And Rosie and I will get along fine … as soon as I break her radio.”

Dr. Irene laughed at that. Letting him go.

Sherlock could tell the man was done with the interview.

But oddly … he wanted more …

“I'm sure you're a wonderful father,” she wrapped up the conversation with a smile, “and even though we’ve only heard and not yet seen your faces … you can tell a lot from a voice.”

“You most certainly can,” John nodded, trying to ring off …

“Just one more question - humour me, John-? Are you sleeping at night?” she asked. Leaving the words hanging in the air …

“He doesn't sleep at all,” little Rosie chimed in from out of nowhere.

“Hey! How do you know that?” John puzzled, looking at his adorably soft child, still rubbing her eyes from sleepiness.

“I live here, Daddy!” she scowled.

They both shared a smile and a giggle as John tickled her and tucked her in closer. Nearly forgetting they were still on the air …

“Look, it’s New Years … I’m tired - I really should be getting her to bed,” he managed at last, “and Mary … my wife … I’ll always miss her. She really was … irreplaceable … “ he sighed, “... and it’s rough this time of year, innit? Maybe Rosie needs someone-? I hadn’t … really thought about that part of things, being wrapped up in my own grief … but I will take it to heart.”

“Do you think … maybe you need someone, too?” Dr. Irene supplied.

It was a rhetorical question,

And they both knew it.

The ending music was now playing, as London’s heartbeat began to pulse again with a new day beginning.

“Well … I’m Dr. Irene Adler,

Signing off with ‘Sleepless in London’ …”

John laughed at his newly christened moniker …

“-and John? Rosie? Will you call in again sometime, soon? And keep us posted on your progress? We’d love to hear from you.”

“We’ll be sure to do that.”

John hung up the phone, quietly. Careful not to disturb Rosie as she had drifted off again to slumberland … still nestled in the crook of his arm.

He hefted her into a more comfortable carrying position, and shifted them both from the window seat to their respective bedrooms …

While Sherlock grabbed his coat and whisked out the door of the lab hurriedly …

Past a stunned-to-silence, and suddenly enlightened Molly … (to be continued ...) Read Part 4

@johnlocky @fluffbyday-smutbynight @chinike @rhasima @chriscalledmesweetie @totallysilvergirl @masterofhounds @missdeliadili @whatnext2020 @safedistancefrombeingsmart @shelleysprometheus @7-percent @swissmissing @inevitably-johnlocked @storytellingdreamer @calaisreno @myriath @ileenhaddockhawkins @discordantwords @john-smiths-jawline @midgemao @raina-at @mxster-jocale @my-johnlockficrecs @anyway-kindness @mydogwatson @gregorovitchworld @kittenmadnessandtea @a-clithridiate-in-my-heart @meetinginsamarra @topsyturvy-turtely @t-dalo @tinchensblog @iwlyanmw @tiverrr @kettykika78 @xeniawatson @sherlockwatsons @janiesprings @copperplatebeech @peanitbear @peageetibbs @cupidford @daltongraham @amyreadsandstresses @pocketwatchofmycroft @kabubsmagga @holmesianlove @colourfulwatson @mrb488 @reveling-in-mayhem

#sleepless in london #johnlock #johnlock fan fic #helloliriels #sleepless in seattle #au #because everything is johnlock #i dont make the rules

72 notes · View notes

githjanken · 8 months ago

Text

so i respect people with a Horny Thing for science but having worked both research and commercial (biochem) labs, i can’t take it seriously. oh yeah, labcoats you say? sex on the workbench, yeah? bestie these people are not going to feel sexy after spending all day fighting over centrifuges*, i booked it, josh, the sign up sheet is there for a reason. trust me.

*substitute appropriate other big machine you only have a limited number of

#seg.misc #let’s not even get started about having to clean the workbench AGAIN if you have sex on them

3 notes · View notes

thessalian · 8 months ago

Text

Thess vs Horizon: Forbidden West Prologue

So a few notes from various bits of prologue. I'm not in the eponymous Forbidden West yet, mostly because one of the things I'm finding is that the low-level anxiety of jumping puzzle hits a little different with the fibro. I get tense, and lemme tell you, I feel it these days. So this is going to be a long, long game. But here are the notes from the Horizon: Forbidden West prologue ... more or less spoiler-free.

Ah, stealth kills, I have missed you. And I've very, very nearly hit the first achievement for those already. Not to mention the "knocking components off" achievement. Not there yet, but ... getting there.

In checking whether I'd actually reached any achivements yet, I note that various friends have got a lot farther than me in this game. Fucking fibro.

See ... this whole deal where it guides you on what actions you need to take to climb or grapple or whatever could be seen as annoying? But I honest-to-gods need a reminder if I'm taking it this slow. So it is annoying but it's also really helpful, so thanks for that.

Some of the new designs on these beasties are lit. Still hate being forced into armed confrontation with these things. Just let me hide and shoot them from the next post code over, damnit!

What's this thing? Can I scavenge it? ...Oh. I was supposed to kill Big Beastie with this. Sorry; I prefer precision. kthxbai.

...Well, if that isn't commentary on tech 'geniuses' and their bullshit, and on corporate greed. Also, it's actually a real shame that so many of the effective antagonists of both games have been dead for so very, very long, because they are all so very, very punchable and apparently they haven't worked out hard light hologram technology so I can't literally punch that smirking shithead in the face.

Aloy, you were raised agnostic and found proof of atheism. Varl has not. Chill the fuck out. You're equipped to see things as they are; your people view everything through a religious filter. So long as they're not exiling people or some shit, why do you care if they worship an AI construct?

Aaaaaaaaaaand we're back to Meridian. Cool. Wait, what the fuck now?

So ... why didn't we double-check the Spire before we went heading off on wild goose chases?

SYLENS YOU DICKWAFFLE. (Sorry, Lance Reddick; you gave really, really good dickwaffle.)

Not sure what to make of this "workbench" mechanic. It'll probably be better when I can fast travel to places.

Please, please tell me I can dump some of the vendor trash while I'm in what passes for civilisation in Horizon-world. Please? I have been so good.

Somebody want to please explain to me why those neckbeard arseholes think Aloy's ugly? Okay, sometimes they do not get her best angle in cutscenes, but mostly she's, like ... a realistically beautiful woman. Ah, right - it's the realistically part they object to. They want doe-eyed skinny waifu. Well, they can get stuffed, because I'm already struggling with how she manages to get the less dreadlocked bits of her hair moving like a shampoo commercial when she has no shampoo. Plus, those neckbeards are not exactly looking at the cover of GQ themselves, y'know?

I'm kind of interested in how they've set up this prologue. Both HZD and HFW start us with where Aloy has been and what she's done, but I guess you can't really turn "killed an invasion of war machines" into an easy prologue the way you can "finding a focus and training for Big Event". Plus, the whole thing makes me wonder exactly how many friendly faces from home we're going to have as companions. I mean, I know two - hell, even if I hadn't seen both in trailers, one I've already had as a companion, and when you get that many Oseram scattered around the place, you know who's going to show up in Horizon-world eventually.

No, seriously, let me offload vendor trash I HAVE BEEN SO GOOD.

So I stopped in and around Meridian before having to tell everybody, "Hey, I'm going to the place named in the title of this game because that guy you told me to look for is a dickwaffle but he could conceivably be a helpful dickwaffle if I help him with the problem he willingly dove into head-first by ... well, being a dickwaffle". I am at least supposed to be finding a workbench and hopefully someone I can throw vendor trash at, but I more or less stopped there because I do have some shit to do today. Not much shit right away - I mean, I do have D&D later this evening and we're still running a little earlier than usual because of Daylight Savings being weird this year, but that's still way later - but some shit. Specifically I need to go out for ground cardamom. Look, I need it to make apple pie spice, and all of my cardamom is in pods. Kind of wish that the recipes for apple pie spice I have looked up specified what colour cardamom needs grinding. I guess I'll just hope I get it right. I need this because I'm baking apple bread. I got buttermilk and everything! Also I want to make more apple turnovers and the homemade filling with only cinnamon ... well, it needs something. Maybe apple pie spice instead of just cinnamon is that something. Anyway, also means I need apples. So that means a shopping trip. So I guess I'd best prep for that.

And then home and more video games. Not sure if I'm going to do my shopping / crafting in HFW or if I'm going to go back to the Shadow-Cursed Lands in BG3 so that I can earn a whole lot of easy approval with Astarion and Lae'zel by making big lumpy bag of poison explode. I'll think about it as I do errands, I guess.

#Thess plays video games #Thess liveblogs Horizon Forbidden West #Thess cooks

3 notes · View notes

nzdepot · 10 days ago

Link

$159.75 $70.29 SOGA 2X 4" Heavy Duty Polyurethane Swivel Castor Wheels - - FC1511WheelsX2 https://nzdepot.co.nz/product/soga-2x-4-heavy-duty-polyurethane-swivel-castor-wheels-fc1511wheelsx2/?feed_id=182616&_unique_id=6739867a010bd Features: Includes 2 fixed casters .Rotate 360 degrees. Smooth & Silent & Sturdy. Loading Capacity: 20-25kg each wheel and 80-90kg for four wheels 4″ Diameter PU Wheel. Polyurethane (PU) wheels offer the elasticity of rubber wheels combined with the toughness and durability of metal wheels. Safe for all floor types Wide range of uses: commonly used in both industrial and residential applications such as workbenches, tool carts, tool chest, shopping carts, cabinets, bookshelves, coffee table and other movable furniture, etc… Specifications: Wheel Size: 99.8cm Screw: 35.2cm x 10.8cm Usage: Home Commercial Package Includes: 2 x Fixed Food Cart Wheels Business […] #

0 notes

toyhaulers · 2 months ago

Text

Why Pick Off-Road Trailers for Your Ensuing Expedition?

For adventure seekers, there are many aids to choosing off-road trailers for your upcoming trip. Because of their harsh terrain construction, these trailers are built to last and remain stable in difficult circumstances. Off-road trailers can handle difficult terrain and isolated areas thanks to features like strong suspension systems, strengthened frames, and high ground clearance. Longer travels are possible without compromising comfort because to their substantial storage capacity for equipment and supplies. The completely camping experience is improved by the fact that off-road trailers frequently come equipped with sleeping quarters and kitchens. Ultimately, for those looking for an extraordinary journey, they are the ideal option.

https://jpcdn.it/img/small/a9379979da8575660e8e57c97754efb5.jpg

Find Every Adventure's Greatest Trailer

Selecting the appropriate trailer is crucial for outdoor exploration and vehicle transportation. The several kinds of trailers that will be covered in this guide are custom tradesman trailers, off-road travel trailers, motorbike trailers for sale, and off-road trailers. These choices all address particular requirements, so you may be sure to have the ideal travel partner.

Sale of Motorcycle Trailers

For any motorcycle lover, it is vital to choose the finest motorcycle trailers for sale. These trailers are made particularly for moving bikes, keeping them safe and secure though preventing damage while in route. Numerous choices are available, from straightforward open designs to Enclosed Car Trailer versions that provide further weather protection. Think about things like loading convenience, weight capacity, and security features when choosing a motorcycle trailer. Investing in a high-quality motorcycle trailer will improve your trip experience and safeguard your car.

Trailers for Off-Road Travel

For individuals who are looking for adventure in difficult terrain, Off Road Travel Trailers are ideal. These trailers are made to withstand harsh surroundings. Their classy suspension systems and sturdy construction provide an even ride across uneven terrain. Off-road travel trailers, furnished through sleeping quarters, kitchens, and restrooms, let you knowledge home comforts although venturing into far-off places. Long treks into the wilderness are made possible by their ample storage for equipment and supplies. Your outdoor excursions can become unforgettable by selecting an Off Road Camper Trailer travel trailer.

Personalized Trailers

Custom trailers provide an individualised solution for those with specific needs. Even if for commercial or leisure use, these trailers can be modified to fit specific needs. Modifications like specific storage spaces, unusual sizes, and extra amenities like integrated toolboxes or workplaces are all possible with custom trailers. You can make sure that a custom trailer fits your lifestyle and improves functionality and convenience while you're on the road by choosing one.

https://jpcdn.it/img/small/f0ee7f691bcf8db65b0694df77d6699f.jpg

Trailers That Are Not on the Road

Adventurers who wish to explore less-traveled paths need Off Road Trailer. These trailers can easily traverse rugged terrain because of their strong construction and exceptional capabilities. All kinds of outdoor activities, from camping to equipment transportation, are appropriate for off-road trailers thanks to features like high ground clearance and heavy-duty tires. Because of their adaptability, they may be used to transport equipment, supplies, or even recreational vehicles, which makes them a great investment for adventure seekers.

Personalized Tradesman Trailers

Custom tradesman trailers specifically meet the needs of professionals in a variety of trades. These mobile workspaces can be configured to hold tools, supplies, and equipment, offering a productive and well-organized work area. Secure storage sections, integrated workbenches, and layouts designed for particular trades are examples of customization options. A specialized tradesman trailer can be purchased to greatly increase productivity by facilitating the transportation of all equipment required for the operation while maintaining security and organization.

0 notes

govindhtech · 3 months ago

Text

CleanML: Optimizing Data And Increasing Efficiency In NER

An MLOps tool for data-centric AI is called CleanML. To assist a machine learning team in managing the Named Entity Recognition (NER) project lifecycle, CleanML was developed. CleanML allows data to be vetted, annotated, modified, and uploaded from a single platform in an effective manner. Annotated data, which can be readily exported in a variety of data formats, may potentially provide insights. Astutic AI, an AI firm that is a part of the Intel Liftoff program, created the tool.

Named Entity Recognition (NER)

The Challenge

It may be quite difficult to oversee the whole lifespan of large-scale, intricate Named Entity Recognition (NER) projects. The procedure entails carefully selecting and annotating the data before drawing the required conclusions from it. This often proves to be an extremely difficult undertaking with the technological tools and platforms available today. This is where artificial intelligence (AI)-based software applications become useful, as they automatically group all connected activities and jobs that need to be finished.

The Solution

CleanML is a SaaS that helps commercial machine learning teams find the best models rapidly. Its major purpose is to improve Named Entity Recognition (NER), an important part of NLP. CleanML unifies key machine learning operations into a single platform, streamlining the analysis and processing of natural language.

The following activities may be accomplished by project managers, data scientists, annotators, and developers using CleanML:

Model training, comparison, experimentation, and lifecycle management

Assessment and correction of data quality

Segmenting data for training and assessment

Sophisticated data retagging and annotating

Using data-centric analytics, CleanML enables teams to find and fix common data and annotation errors while simultaneously maintaining and experimenting with models to optimize performance. Additionally, you can monitor and compare training iterations and see how modifications to data or code affect overall model metrics by delving deeply into model evaluation records.

Who stands to gain from CleanML use?

Project managers have the ability to start many projects and monitor each one separately.

Data scientists may learn more about the distribution of annotated entities, training and test data, and how to curate additional data for greater accuracy.

With CleanML‘s useful features, annotators may expedite and enhance the annotation process from a single window.

Using various libraries, software developers may test various algorithms on GPUs, CPUs, on-premises systems, or cloud computing environments.

CleanML‘s data-centric dashboard makes it possible to identify and resolve problems with data and data categorization. Additionally, drill-down analytics on the dataset are possible.

Which are the salient characteristics?

Data-centric dashboard: Investigate in-depth analytics while identifying and fixing data and categorization issues. Learn about the grouping of data and what categories were omitted or out of the ordinary.

Advanced workbench: Access to prior classifications, text annotation, renaming entities across records, in-place content editing, tag and auto-labeling recommendations, and the ability to build a custom dictionary are just a few of the useful capabilities that Workbench provides.

Built-in data versioning: CleanML keeps track of data versions automatically, which facilitates the replication of training outcomes. It also lets you evaluate how well a model performs when compared to other models, versions, and even methods that are presently being used.

Train, test, compare, and repeat: Using the same dataset, train and compare models using various techniques. CleanML keeps track of training and data versions, which facilitates in-depth comparisons and increases efficiency.

Auto-labeling recommendations: To expedite and simplify data annotation, get automatic labeling ideas from algorithms that have been trained.

Advanced Workbenches from CleanML provide practical capabilities like text annotation, entity renaming across records, in-place content editing, and more.

Models using several techniques may be trained and compared using the same dataset.

How does CleanML assist?

CleanML versions all of the training and facilitates data and training version comparisons.

Scientists of Data: Learn more about the distribution of annotated entities, training and test data, and how to curate additional data for greater accuracy.

Annotators: Uses CleanML’s useful features to enhance and speed up the annotation process all from a single window.

Developers: Try out several algorithms with various libraries, whether they be on a GPU, CPU, on-premises, or cloud.

By training a new model or algorithm in CleanML and comparing its results to those of the existing models in your project, you may quickly experiment with it.

Learn many algorithms, then contrast them.

Monitor the data annotation process.

Develop unique word embeddings for applications that are domain-specific.

Examine each entity’s over- and under-fitting in relation to the training results.

Incorporate production data and evaluate the model’s accuracy in the real world.

Compare the production model to a recently created and trained model.

Features

Dashboard focused on data

Determine and address problems with the data and data categorization, and do drill-down analysis on the dataset. Learn about data that has been categorized into many categories or classes, missing classifications, and classification anomalies.

Advanced Workstation

A custom dictionary may be added to Workbench, along with other helpful capabilities like annotation of text, entity renaming across records, in-place content editing, tag recommendations, auto-labeling suggestions, and prior classifications.

Integrated data versioning

Data versioning is done automatically by CleanML. Reproducible training is aided by this. Additionally, CleanML offers the option to compare a model’s training with that of an upgraded version of the model, a model that employs an alternative method, and even a model that has been put into operation.

Test, Train, Compare, and Repeat

Models using several techniques may be trained and compared using the same dataset. CleanML versions all of the training and facilitates data and training version comparisons. Your productivity will rise dramatically if you can compare both models and data at a record level.

Automatic labeling recommendations

Receive recommendations for labeling based on algorithms that have been learned; they may help annotators and expedite the annotation of fresh data.

Multiple data types are supported

Data may be imported into CoNLL-2003, JSONL, txt, IOB (IOB1/2, BILOU, IOBES), and IOB. Bring in data from the Singer Taps, UI, API, and command line. Utilize the command-line to export annotated data in a variety of data formats.