Why Russian Drones, Missiles' Lower Quality Raises Vulnerability To Ukraine EW

Fewer Russian drones and missiles are getting through, either because they are shot down or because they simply fail to reach their targets. 

The root cause is the same: as it becomes harder for Russia to source electronic components on the gray market, the quality of their drones and missiles has decreased. This means they either fail outright - or they become easier for Ukrainian electronic warfare systems to target. JL 

The New Voice of Ukraine reports:

Russia launched 40 cruise and ballistic missiles at Ukraine on Jan. 13. 8 of them were shot down, while 24 exploded before reaching their targets due to Ukrainian electronic warfare. The quality of Russian missiles is declining as reducing quality is a way to boost production in Russia. This means systems become more vulnerable to EW. Some missiles are no longer equipped with complex guidance systems for cost-cutting reasons. Quality has declined to the point that it doesn’t work properly. The missile can be confused and go off course to avoid flying where it has a link to trajectory correction systems. A signal guides the missile towards what it “sees.” It’s possible to affect such homing mechanisms.

Ukraine's 3rd Assault Brigade Using Mine Clearance Ground Drone

The use of such unmanned vehicles is growing, in order to protect troops' lives and to reduce obstacles that might otherwise be too difficult to overcome. JL 

Militarnyi reports:

A ground drone for clearing tracks of land mines for assault groups was developed in Ukraine. It has been tested by Ukraine's 3rd Assault Brigade. These drones operate at close distances to Russian positions that the Ukrainian infantry is going to storm. The drone was tested with PMN-2, PMN-4 and OZM-72 anti-personnel mines. The drone also detonated a TM-62 anti-tank mine.

Tests Reveal How AI Violates Copyrighted Material That Is Not Fair Use

Tests reveal AI routinely violates copyrights and other intellectual property by changing minor details. 

AI companies are claiming 'fair use' protects their using images for which they dont have permission and for which they havent paid, but new legal cases are challenging that assumption - and may threaten the exploitative AI financial model. JL 

Stuart Thompson reports in the New York Times:

A.I. companies violate copyright in two ways: train on copyrighted material they have not licensed, or reproduce copyrighted material when users enter a prompt. A.I. companies’ use of intellectual property raise questions about the training data used to create every A.I. system and whether the companies are violating copyright laws. When journalists asked ChatGPT to create an image of SpongeBob SquarePants, it produced an image similar to the cartoon. The differences were that the character’s tie was yellow instead of red, and it had eyebrows instead of eyelashes. When journalists omitted SpongeBob’s name from a request, OpenAI created a character that was even closer to the copyrighted work.

US and NATO Long Term Ukraine Strategy Focused On 'Russia-Proofing' Defense

It is also designed to Trump-proof the US defense budget in order to prevent any further Putin influence on Republican decision-making which is contrary to long term US strategic interests. JL 

Karen De Young and colleagues report in the Washington Post:

A multilateral effort by nearly three dozen countries backing Ukraine to pledge long-term security and economic support is to position Ukraine to hold its position on the battlefield for now, but “put them on a trajectory to be much stronger by the end of 2024 and get them on a sustainable path” to concentrate where its forces have had recent success - longer-distance fires, including cruise missiles; holding back Russia’s Black Sea Fleet; tying up Russian forces inside Crimea with missile strikes and sabotage; and air defense to protect Ukrainian cities and rebuild the Ukrainian economy. The long-term promise will also “future-proof” aid for Ukraine against the possibility that Trump wins.

How Ukraine's "Drone Geeks" Have Fought the Russian Army To A Standstill

Using the same skillsets, organizational models - and brilliant, quirky individuals - a dynamic cottage industry has emerged in Ukraine to build increasingly effective and lethal drones for the Ukrainian military. 

The result is that despite shortages of parts, money and engineering degrees, the Ukrainians have created an industry whose products and the people who pilot them have fought the ostensibly superior Russian army to a standstill after two years of war. JL 

Alya Shandra reports in Euromaidan Press:

FPV drones, used for racing, stream video directly into the pilot’s goggles. For $500, an FPV drone can uncover a dugout so mortars can blast it, take out a T-72, and set a Russian Tor air defense system ablaze. “These are expensive, irreplaceable losses for Russia." Many little garage operations produce their own for different tasks. Ukraine has dozens of such groups, and (some) in each brigade. Agility and flexibility are key. They have direct feedback from the units they supply, allowing rapid problem solving. “You don’t need a big plant and Ivy League engineers. Combining financing, enthusiasm, and knowledge works.” Anyone capable of handling a soldering iron can be taught drone construction.

The Reason Avdiivka Became A Junkyard Of Destroyed Russian Armor

In scenes reminiscent of Ukrainian forces destroying attacking Russian armor at the outset of the invasion in February 2022, the Ukrainians are again decimating Russian armor attempting to advance around Avdiivka. 

The Russians appear to be following predetermined routes, expecting that overwhelming force will deliver the result they seek. But Ukrainians increasingly effective use of drones for sighting and targeting Russian armor have rendered that Stalinist brute force tactic less effective than it was 80 years ago resulting in massive Russian losses. JL 

Chris York reports in the Kyiv Post:

Geo-located footage shows dozens of burnt out, destroyed Russian vehicles along the route into Stepove, north of Avdiivka. "This is just one of the intersections of forest belts in the Avdiivka area, and there are dozens of such places.” During the battle for Avdiivka Moscow has lost 574 vehicles while Kyiv has only lost 44. Ukraine’s 110th Mechanised Brigade shows Russian armored columns decimated (by) Ukrainian artillery. “Ukrainian counterattacks are holding Russian forces from progressing. As the main supply route remains intact, and Ukrainian forces make local counterattacks, Avdiivka is likely to remain in Ukrainian control over the coming weeks.”

Ukrainian Bradley That Took Out Russian Tank Had Crew Of Only Two

The Bradley is not as well armed or armored as a Russian T-90 tank. But the story that a Ukrainian Bradley crew took out a T-90 has made global headlines.

As if that news werent dramatic enough, Ukraine has now revealed that the Bradley - designed for a crew of at least 3 - was that day fighting with only a crew of 2 men - a driver and a gunner, making its victory even more improbable. JL 

Nicholas Slayton reports in Task and Purpose:

Not only was the Ukrainian Bradley outgunned and up against a Russian unit with better armor, it was likely doing so with a smaller number of soldiers. The general rule for armor is at least a crew of three: a driver, a gunner and a commander who can spot targets. It allows each crewmember to focus on their primary tasks in the most efficient way. That’s true for how the Bradley Fighting Vehicle is meant to be used. It’s not clear why the Ukrainian Bradley was only using a two-person crew, or if that is a common practice among Ukraine's 47th Mechanized Brigade

So Few Eligible Russian Convicts Remain, Kremlin No Longer Offering Pardons

So many Russian convicts have been coerced to fight in Ukraine have been killed that there are very few left in the Russian prison system physically or psychologically eligible to do so. 

As a result, the Kremlin is no longer offering pardons, short enlistment periods and has reduced the financial benefits to convicted felons because it no longer can use large numbers of them. This also helps reduce the war's burden on the Russian government budget. Approximately 54,000 convicts were recruited, the vast majority of whom are presumed dead, killed in fruitless 'meat attacks' known for their suicidal nature and limited survival rates. JL 

Institute for the Study of War reports:

The Kremlin is no longer offering pardons to convict recruits and is significantly changing the terms of their service in response to the reduction of the pool of convicts suitable for Russian forces. As of October 2023 the Russian prison population was 266,000 people, a notable reduction of 54,000 prisoners from January 2023. The loss of convict recruits to attritional assaults in Ukraine and the short terms of their service contracts prompted the Kremlin to enact more restrictive terms of service in order to retain more convict recruits at the front in Ukraine and to relieve force generation burdens on the Russian federal budget.

Why Even At Peak Advantage, Russia Is Unable To Achieve War Objectives

Russia's army in Ukraine may never be stronger than it is right now, with new ammunition from Iran and North Korea, as well as enough troops. 

 But despite its advantages - heightened by Russian-influenced blocking of US and EU aid - Russia has proven incapable of achieving Putin's objectives for this war. The reasons reveal Russia's weaknesses: the relative inferiority of its weapons, poor leadership and appalling tactical decision-making which may have worked for Stalin in 1943, but is out of date in the current environment. The result is that if Russia cant figure out how to win now, it may never do so. JL 

Michael Kofman and colleagues report in War On the Rocks:

In 2024, the best defense is not likely to be a good offense, but rather one that maximizes efficiency and creates the right opportunities down the line. Ukraine can play to its advantages while defending, leveraging improved long-range strike capabilities to target Russian critical infrastructure far behind enemy lines. The active component of the strategy is an extended strike campaign that helps set the conditions in 2025. Russian leadership is now visibly overconfident. Even at the peak of its defense spending, Russia is still unable to achieve its objectives in this war. The costs will mount, and it is Moscow that will face growing uncertainty in 2025

Ukraine's Drone Targeting Is Preventing Russians Troops Even Reaching the Front

Ukraine's increasingly effective use of drones to identify, target and destroy Russian logistics, armor and troop formations has so frustrated Russian army leadership that they can no longer count on troops even getting to the front. 

The result is a battlefield stalemate at a time when Russia enjoys superiority in troop strength, artillery, missiles and armored vehicles. JL 

Tom Porter reports in Business Insider:

Ukraine's targeting of Russian troops and armored vehicles is credited with "preventing Russian forces from even reaching Ukrainian forward defensive lines." Ukraine has excelled in the use of cheap hobby drones on the battlefield, using them to surveil Russian positions and fitting them with explosives. Russian forces were "unable to concentrate in numbers sufficient to break through Ukrainian lines because Ukrainian forces strike all force concentrations larger than a battalion. The more abundant, smaller drones are proving to be serious game changers in that they have given Ukraine better battlespace awareness and more capability to hit targets,"

Ukrainian HIMARS Kills 24 Russian Drone Pilots Gathered For Training

Experienced drone pilots had gathered to improve their skills and train new recruits. 

Indiscreet notice of the gathering gave Ukrainian drone and HIMARS operators the opportunity to eliminate them. JL 

Isabel Van Brugen reports in Newsweek:

Kyiv's forces attacked a Russian military training ground in Ukraine's eastern Donetsk region on Wednesday, striking a group of drone pilots. The drone operators were struck near the city of Ilovaisk because the head of the Sudoplatov (or "Judgment Day") group carelessly gave out coordinates on its positions to "strangers" via a bot on "where to come for training." Sudoplatov is a Russian volunteer group that produces 1,000 drones per day. "Cadets are trained, and experienced fighters arrive to improve their qualifications, as well as transfer their combat experience gained on the line of contact."

AI Will Transform the Global Economy, But Will It Benefit Humanity?

The issue most of those concerned about AI's impact focus on is that it will 'steal' human's jobs. 

But as more economic impact research is conducted, the real problem may not be loss of jobs, but on worsening inequality between those countries, businesses and individuals who learn how to effectively use it and those who do not. The challenge, in that scenario, is what policies can be implemented to assure broader distribution of its benefit for the broadest positive outcome. JL 

Kristalina Georgieva reports in the International Monetary Fund:

In most scenarios, AI will likely worsen overall inequality. In advanced economies, 60% of jobs may be impacted by AI. Half may benefit from AI integration, enhancing productivity. For the other half, AI applications may execute tasks currently performed by humans, which could lower labor demand, leading to lower wages and reduced hiring. AI could also affect income and wealth inequality within countries. We may see polarization within income brackets. If AI complements higher-income workers, it may lead to a disproportionate increase in their labor income. Gains in productivity from firms that adopt AI will boost capital returns, which may also favor high earners, exacerbating inequality.

The Graveyard of Command Posts: Lessons In Ukraine's High Tech Trench Warfare

Mobile, fast, plentiful, miniaturized, cheap and well hidden. 

The lessons of high tech trench warfare in Ukraine, which combines elements of the Jetsons and of World War I are that modern electronics and drones offer few means of escape. But that old school camoflage and digging in can mitigate the effects. In addition, the need for quantity over quality will remain a point of debate - but quantity seems to have the upper hand in some areas (drones), quality in others like Bradley fighting vehicles, smart munitions - and well trained troops. JL

Sydney Freedberg reports in Breaking Defense:

The Ukraine war has shown that big, static targets do not last long against a well-armed foe with their own drones and long-range artillery. Ideas range from moving non-essential data from bulky “tactical servers” to the cloud, to old-school soldiering tricks, like hiding your command post in the basement of a rubbled building. Dumb and cheap: When facing electronic warfare in Ukraine, small drones’ quantity is quality. It is more cost-effective to buy cheap drones in bulk and treat them as disposable, than try to upgrade them with better defenses against electronic warfare, let alone physical gunfire. Let units across land, sea, air, outer space and cyberspace share targeting data and coordinate strikes in near-real-time. In cyberwar, prepare for prolonged conflict, not a knockout blow

Ukraine Blows Up Another Russian Oil Refinery - Its Fourth In Six Days

That's a lot of fuel for tanks, armored vehicles, logistics support - and heating - to say nothing of earning export income. 

And four in six days suggests they are not especially well-defended. JL 

Stefan Korshak and Kateryna Zakharchenko report in the Kyiv Post:

Explosions rocked Russia’s biggest oil refinery on the Black Sea late Wednesday evening, Tuapse oil refinery, operated by Russian oil company Rosneft. It's the major fuel and lubricants processing facility on the Black Sea, in pre-war years producing two million tons of product a year, almost all of it for export. In the past six days, Russian oil and gas processing infrastructure has been hit with major drone raids four times. One drone bound towards Tuapse apparently traversed portions of Russia’s densely populated Sochi region beach resort region, cruising at 200-300 meters’ altitude and easily visible to residents below.

The Reason Amazon Ending Police Access To Ring Home Surveillance Video

Ring denying automatic police access to its data  may, secondarily, be a response to growing consumer concern about privacy. 

But the likelier reason is that Ring is rethinking its market with new hardware and services, so likely wants to increase the amount of data it captures but in a way that is more monetizable for the company; in other words, no one - including the police - gets anything for free. JL 

Matt Day reports in Bloomberg:

Amazon’s Ring home doorbell unit says it will stop letting police departments request footage from users’ video doorbells and surveillance cameras. Police and fire departments will have to seek a warrant to request footage from users or show evidence of an ongoing emergency. Ring decided to devote its resources to new products and experiences in the Neighbors app. The aim is to make Neighbors, which had been focused on crime and safety, into more of a community hub. Ring is rethinking its mission, in part to account for its expanded portfolio of devices

Ukraine Retains Avdiivka Control After Pushing Out Russian Recon Unit

Russia's claims of a 'breakthrough' at Avdiivka turn out to be premature. 

Ukrainian forces report that Russian recon and sabotage teams briefly entered the southern part of Avdiivka but were quickly pushed out and that Ukraine retains control of the city. JL 

Jon Jackson reports in Newsweek:

Russian forces recently entered Avdiivka for the first time but were promptly pushed back. "Russian sabotage and reconnaissance groups entered the southern part of the city of Avdiivka, but they were dislodged." Institute for the Study of War "stated that positional fighting continued" near Avdiivka. The U.S.-based think tank said geolocated footage indicated Russian forces advanced in some nearby settlements such as Stepove. "The situation is under the control of Ukrainian armed forces, difficult but controlled."

Why Russian Transport Plane Shoot Down Looks Suspicious, Like Disinformation

There are several equally unsavory explanations for the shooting down of the Russian transport plane. One is that it contained missiles, as Ukraine claims, but not Ukrainian prisoners and that the Russians are simply making it up to try to deny the destruction of more planes and weapons while sowing discord.

A second is that the Russians intentionally placed Ukrainian prisoners on the plane with the weapons as human shields but did not provide the usual notification not to shoot over that route, knowing the Ukrainians might shoot it down, giving them a propaganda point while killing Ukrainians. There are reports that Russian VIPs scheduled to accompany the missiles were prevented by the FSB from boarding the plane at the last minute, which would further confirm its use as a trap. So far there have been no pictures of bodies beyond those of the plane's crew - and rescue workers have been kept away from the crash site. JL

Valentyna Romanenko reports in Ukraine Pravda:

Defence Intelligence of Ukraine says that the Il-76 plane that crashed in Russia’s Belgorod Oblast the day before was supposed to carry Russian officials, but at the last minute, the FSB forbade them to board. A prisoner swap had been due to take place on Wednesday, (but Ukraine did not received customary notification of their transport over that area). After the incident, only five bodies - the number of crew members of the Il-76  - were delivered to the local morgue. Video from the crash site does not show human remains. Rescue workers were not allowed to inspect the crash site. Ukraine says the aircraft was a military target and was carrying ammunition (missiles) for the Russian army.

In 700 Days Since Russia's Ukraine Invasion Its Defeat Is Still Closer Than Victory

Russia forces remain mired in their own incompetence and ineffectiveness with little to show for the @370,000 dead and incapacitated Russian soldiers who've paid for Putin's misadventure with their lives and futures. JL 

Mark Sumner reports in Daily Kos:

Wednesday marked the 700th day since Russian tanks rolled into Ukraine. Military analysts around the world predicted Russia’s conquest of Ukraine would look like America’s shock-and-awe campaign in Iraq. Almost everyone expected Ukraine to lose quickly. Except the Ukrainians. Officers in that initial “40-kilometer convoy” were told to bring dress uniforms for military parades in Kyiv. The next time anyone saw those Russian dress uniforms, they were being pulled out of destroyed tanks. Putin couldn’t take Kyiv in three days. Or three weeks. Or three months. And he’s not going to take it in three years.

Ukrainian Hackers Stole Construction Plans For 500 Russian Military Sites

The Ukrainian hackers not only stole the documents but deleted the servers holding the backups. JL

Rebecca Rommen reports in Business Insider:

"Blackjack," a Ukrainian group of hackers with possible ties to the country's main spy agency, stole construction plans for over 500 Russian military sites. A successful cyberattack was launched against the Russian state enterprise overseeing all construction contracts for Russia's Ministry of Defence. The Blackjack group amassed 1.2 terabytes of classified data on Vladimir Putin's military apparatus. The data includes detailed maps of more than 500 Russian military bases across Russia and Russian-occupied Ukrainian territories. The group also took down seven servers and encrypted over 150 of the Russian contractor's employee computers.

How Impact of Ukraine Rear Area Strikes Confine Russians To Positional War

Internal Russian reports reveal frustration at their inability to stop devastating Ukrainian strikes on Russia's rear area and logistics which has degraded capabilities as well as reduced Russian ability to attack.

The Russians are unable to concentrate sufficient forces or materiel without being seen and hit by Ukrainian drones and missiles. The result is that they are forced into positional warfare despite their superiority in troop strength and weaponry. JL

Critical Threats reports:

Russian forces are struggling to compensate for Ukrainian drone and rear-area strikes at the level necessary to break out of positional warfare. They are unable to concentrate in numbers sufficient to break through Ukrainian lines because Ukrainian forces strike all force concentrations larger than a battalion. Ukrainian forces target small Russian groups of one-to-two infantry companies and of 10 armored vehicles with drone strikes, preventing Russian forces from even reaching Ukrainian forward defensive lines. Russians have (also) complained about poor Russian indirect fire,  attrition through unproductive “meat assaults”, poor tactical and operational planning, and countering drone operations. There are no indications the Russian command has improved on any of these issues

Lessons Learned From Ukraine's Patriot Missile Experience Prove Invaluable

This is a big part of the reason why Ukrainian air defense intercept rates are continuing to improve. A benefit from which its NATO allies will also gain advantage. JL

Jan Kallberg reports in the Center For European Policy Analysis:

When Ukrainian Patriot missile air defenses repeatedly shoot down the Russian Kh-47M2 Kinzhal, a missile the Kremlin describes as a wonder-weapon it not only saves the military and civilians from damage, it provides the ultimate test range. That’s because Patriots — Ukraine’s only defense against ballistic missiles — and multiple other weapons systems are being trialed against a peer rival in real time. For all the improvements in computer modeling, the real battlefield remains unmatchable. When the Ukrainians intercept a Russian missile, bomber or command plane, the data is fed back into the Patriot missiles intercept algorithm, teaching the air defense system to become better over time.

How Ukraine's M1A1 Abrams Tanks Make Other Armor More Effective

The M1's sensors, vision range, thermal acuity and digital zoom function gives it more intelligence and accuracy than any other tank operating on either side in Ukraine.

With this capability, it makes the other armored vehicles around it - and the infantry they are supporting - more deadly at a greater distance, thereby improving the effectiveness of an entire unit, not just itself. JL

David Axe reports in Forbes:

Ukraine’s 31 US-made M-1 Abrams enhance interoperability. With its thermal vision system, a four-person M-1 can detect targets up to five miles away—farther than all other main battle tanks in Russia's war on Ukraine. An M1 gunner should be able to identify it, too, thanks to the 50x digital zoom function. Pairing an Abrams with other vehicles also lends tremendous night-fighting capability. The M-1 high-fidelity thermal sights fights almost as well at night as it does during the day, out to several miles compared to the Russian T-72, whose thermal sights might work at a little over a mile. The M-1 enhances the situational awareness of other vehicle types and help them to fight around the clock.

Russia Is Losing 400 Men Per KM, Despite 6 Times More Artillery Shells

The Ukrainians continue to fight smarter than the Russians, which is why their casualties are lower while the damage they are inflicting on the Russians is greater. JL 

Orysia Hrudka reports in Euromaidan Press:

Russia is currently losing approximately 400 soldiers for every square kilometer of Ukrainian territory captured. Additionally, Russia used six times more artillery shells than Ukrainian forces. Ukraine’s air defense capabilities had improved significantly (despite)“In the last two months, (Russia) unleashed over 600 rockets and more than 1000 Shahed drones on Ukrainian cities. The number of Shahed drones launched increased by over a third.”

Google Cuts XLab Jobs, Wants 'Moonshots' To Find 'Market-Based Capital'

Even with the economy stabilizing, interest rates starting to come down and AI interest (if not profits) growing, Google/Alphabet has determined that it just cant continue to burn $1 billion a quarter on internal 'moonshot' startups. 

Instead, it wants them to seek outside venture funding as proof of viability using other peoples' money. And before asking who might want Google's cast-offs, there is a history of former Google XLab startups going on to greater success, so this could be an opportunity for the right fund. JL 

Ron Amadeo reports in ars technica:

The X Lab is Alphabet's "moonshot" group, responsible for wearable display, a self-driving car, smart contact lenses, Internet balloons, and delivery drones. None of those projects is a commercial success. X Lab is part of Alphabet's "Other Bets" group, which burns a billion dollars every quarter. "We’re spinning out more projects as independent companies funded through market-based capital to collaborate with industry and financial partners." Google wants these money losers to find their own funding. Outside funding isn't new for Alphabet. Waymo took outside funding, racking up $5 billion. Alphabet's health data analytics company also raised billions outside. Both "graduated" to full-fledged Alphabet companies. Startups within X often face waiting for a spot to open up or striking out on their own.

Why Has Russia's Ukraine Commander Gerasimov Been Missing For 2 Weeks?

There are two theories as to his disappearance. One is that he was killed - either purposely or inadvertently - by Ukrainian missile attacks on Russian bases in Crimea. This holds that either the Ukrainians knew he was there and targeted him - or that he just happened to be in the wrong place at the wrong time. The other theory is that Putin lost patience with him for not having delivered a victory and has had him eliminated. 

Russia watchers say that if the Kremlin eventually says he died of a medical condition, then Ukraine probably killed him. But if they say Ukraine killed him, the Kremlin probably did so. He's too important to have 'fallen out of a window' as so many other high ranking Russians have done but either way, he hasn't been seen and no one's talking. JL

Jonathan Sweet and Mark Toth report in the Kyiv Post:

Gerasimov’s absence from the public and Moscow’s silence on his status are interesting. 23 Russian troops were killed in the Ukrainian attack on the Russian airbase in Saky, Crimea. Five were reported to be “high-ranking commanders.” Rumors quickly spread that one was Gerasimov. He was last seen in public was December 29th. His efforts have failed to deliver the knockout blow Putin demanded. Worse, Russian forces were turned back in Bakhmut and Avdiivka at an abysmal cost, while the Black Sea Fleet was forced to abandon their headquarters in Sevastopol. (But) his doctrine has certain degrees of success. Particularly escalating conflict via proxies in other regions including the Middle East.

As Russian Casualties Soar, Dead, Wounded Are Left Frozen On Ground

Russian casualties are soaring according to NATO analysts. There are so many dead and wounded - and the temperatures are so cold on a battlefield made extremely dangerous by drones - that the Russians are not even attempting to retrieve the bodies of their fallen anymore. JL 

Sinead Baker reports in Business Insider:

Russia's "meat assault" tactics are leaving frozen bodies that Russia isn't collecting. New Russian troops often receive little training, according to Western intelligence and captured Russian soldiers. Families of Russian soldiers claimed last month that wounded troops were being sent back to Avdiivka, and that it was tantamount to their deliberate extermination. "If we kill 40 to 70 servicemen with drones in a day, the next day they renew their forces and continue to attack."

What Happened When AI Drafted A Startup Business Plan

AI is able to provide a basic, 'plain vanilla' business plan - but was unable to offer any of the detail and personalization about markets, finance and personnel that differentiate between successful businesses and those which fail to ever get off the ground. 

It is, in short, a useful, if limited template that requires either greater development - or experienced, intelligent human intervention. JL 

Gordon Fletcher and Maria Kutar report in the Wall Street Journal:

What did we find? In short: The business plans produced by AI were useful, but not quite ready for prime time. They could serve as a good starting point for a company founder, but are in need of further editing and additions. The plan contained all the sections expected and provided a solid foundation. AI was good at identifying critical elements, such as unique selling points. (But) AI isn’t going to win any awards for creativity. It repeated certain phrases and concepts; fell short in  market data, financial projections and business challenges; contained occasional inaccuracies (AI 'hallucination'); and it struggled to personalize or adapt the structure to the business being described.

How A Russian Attack On NATO Might Proceed Based On Ukraine Experience

This is a hypothetical scenario - one of many that western militaries explore regularly in order to prepare for potential threats. 

It is not necessarily likely - or even possible - but reflects current expectations based on Russia's behavior in Ukraine. JL 

David Averre and colleagues report in Mailonline:

NATO countries are gearing up for one of their largest-ever drills, with 90,000 troops from all corners of the continent and the US set to embark upon war games. In the first phase, cyber attacks disrupt logistics and supply chains; conflict in space, with satellites used to attack each other or jamming conducted against satellites; disinformation campaigns; long-range precision missiles against civilian targets all over Europe. A second phase would involve sending thousands of Russian soldiers, AI-controlled tanks and special forces, to attack the Baltic states as the Russian Navy takes control of the Arctic. Iran, Hamas, Hezbollah, Yemen's Houthi rebels, Iraq, Syria, North Korea and China could join Russia.

Ukrainians Amazingly Down Cruise Missile With Browning Machine Gun

Luck, practice or random numbers. It is theoretically possible to do. Whatever works. JL 

Ellie Cook reports in Newsweek, image Anatolii Stepanov, AFP via Getty:

Ukraine intercepted a Russian cruise missile by using a Browning heavy machine gun. "It is certainly possible to shoot down a cruise missile with a machine gun," although it is very difficult to do in practice. "A cruise missile is a small, very slow-flying jet aircraft with a lot of explosive on board and a single bullet can easily cause catastrophic damage. The hardest part is actually hitting the incoming cruise missile from a distance as it travels at hundreds of miles per hour. If you have enough machine guns putting enough lead into the air and enough missile targets, sooner or later, someone will get lucky."

How Ukraine Can Achieve Conventional Warfare Victory With Enough Drones

Much of Ukraine's battlefield success to date can be attributed to its creative use of drone warfare against an enemy superior in size and resources. 

Drones have made it possible for Ukraine to hold its own even as European countries build up capacity - and as the US battles Russian corruption of some of its politicians. With greater range, accuracy - and lethality - drones can be expected to help degrade Russian capabilities going forward. JL 

Peter Olandt reports in Daily Kos:

The challenge of the Ukrainian battlefield is that the attacking infantry is observed by a small drone from kilometers away which calls in artillery fire (or drones). We have evidence that Ukraine can locally control electronic warfare to limit Russian drones. One tactic is targeting enemy drone operators and transmitters, destroying them.  Russian drone operators are less willing to go near the front now. So the drone problem may be solvable using old military concepts. Maneuver warfare can result in a smaller force defeating a larger one as the smaller force avoids the main mass of the enemy while striking at critical points causing the collapse of the large force. Ukraine can accomplish this.

Why Russia's St Petersburg Air Defenses Cant Protect Against Ukraine Attacks

Russia's air defenses in the St Petersburg region were designed to counter air attacks by NATO - which would come from the northwest or west. Ukraine's attacks are coming from the south, making it easier for them to evade detection. 

Russia may now be increasing defenses around St Petersburg as Ukrainian attacks have severely damaged to two oil and gas facilities. But to do so, they are going to have to take additional air defense assets away from the border, giving Ukraine additional attack opportunities. JL 

Ukrinform reports:

Russian air defense cannot effectively protect the airspace over the St Petersburg -Leningrad region, especially attacks coming from the south, as it is configured exclusively to repel hypothetical NATO attacks which would originate from the northwest and west. Russia has historically built up its air defenses in the area to defend against NATO. "Ukrainian strikes in Leningrad Oblast may (now) prompt Russian forces to reposition short-range air defense systems along expected flight routes of Ukrainian drones to defend potential targets of strategic value."

Ukraine Bradley That Beat Russian T-90 Tank Shows Crew Skill Is Crucial

The Ukrainian M-2 Bradley infantry fighting vehicle crew that destroyed a state of the art Russian T-90 tanks had returned from training in Germany two weeks earlier. They ran out of armor piercing ammo. But rather than turn and run, they applied their training and skill, targeting the tank's optics. 

Unable to see, the tank spun around, hit a tree and the crew bailed. A Ukrainian drone then finished it off. Skills matter. JL 

David Axe reports in Forbes:

All things were not equal in Stepove that day. A T-90’s 125-millimeter smoothbore main gun makes the tank more than a match for an M-2 with its smaller gun and much thinner armor. Unable to penetrate the 51-ton T-90, (the Ukrainian Bradley gunner) targeted the tank’s fragile optics. “I started blinding him so he couldn’t leave.” Hammering the T-90 with one-pound autocannon rounds triggered the tank’s reactive armor and destroyed the optics. Its turret spinning, the tank rolled out of control—and into a tree. The three crew bailed out. The driver got captured. Later, a Ukrainian FPV drone finished the T-90. Its wreck still was on the battlefield days later. Skill matters as much as equipment.

Ukraine Is Building Its Own Surovikin Line. And It's Lethally Daunting

Ukraine keeps learning and adapting - but always with a new twist. Rather than simply dig in, Ukraine is expanding its assault on Russian economic targets and logistics to weaken Russia's military capabilities while continuing to probe for weak points into which it can launch attacks. JL 

Joe Barnes reports in The Telegraph:

Rows of freshly-dug trenches, concrete dragon’s teeth and underground command centres adorn vast expanses of Ukraine’s countryside. The new fortifications have sprung along key segments of its more than 621 miles of front (see map in article below). The new defences resemble the “Surovikin line” three-layered system of trenches, tank traps and strong-points Russia used to thwart Ukraine’s counter-offensive. Minefields that plagued Ukraine’s NATO-trained brigades have been replicated by Kyiv. (But rather than static) “active defence” holds lines while maintaining offensive action to find weak spots, similar to the surprise offensive in September 2022 that saw Kyiv liberate the Kharkiv region.

Ukraine Drone Skills give Russia's Cross-Dnipro Troops Have No Place To Hide

Ukraine's drones - and their growing number of pilots have made it difficult for Russia to mass troops to attack Ukrainian positions, as well as to resupply or rotate in replacements. 

The devastation is sufficient to have prevented the Russians, who outnumber the Ukrainians 4 to 1 in the cross-Dnipro sector, from pushing them out of their bridgehead. JL 

James Waterhouse reports in the BBC:

Ukrainian units target a main road which the Russians use to deliver supplies. Civilians are banned from driving there, so the Ukrainian drone pilots hit anything with wheels. Drones mean the Russians can't hide anywhere within 10km (six miles) of the front. On the outskirts of Kherson in an icy field, pilots practise drone flights with plastic bottles beneath them, in place of grenades. It takes14 hours of training to qualify as a drone pilot. Ukraine encourages people to take free training, as well as manufacture drones at home. "We are engaged in a struggle of technologies, an arms race. Our goal is to destroy their pilots. We have the coordinates, so we're flying there."

The Reason AI Is Becoming One of the Keys To Ukranian Military Success

Ukraine is using homegrown and western AI for an increasing number of military tasks, including target location and range, drone direction - and identification of individual Russian soldiers.

Its application is effective and enhances the power and reach of the undermanned and gunned Ukrainian army. JL 

Euhenia Martinyuk reports in Euromaidan Press:

Ukraine uses western AI developments for controlling drones and targeting; collecting evidence of war crimes and locating war criminals; detecting Russian disinformation and propaganda; demining; and reconstruction planning. With just one photo of a Russian soldier, Clearview's facial recognition AI revealed name, hometown, and social media profiles, even with grainy, partially obscured images. It has helped them identify 230,000 Russian soldiers. "Ukraine is building it from the bottom up, and it’s antifragile … it’s small, it’s scalable, it works, and they know what to do with it."

How Generative AI May Boost Productivity Without Replacing Workers

New research suggests that rather than replacing workers, generative AI chatbot assistants may help inexperienced or less skilled employees become considerably more productive by assisting them with better knowledge and tactics. 

When scaled this knowledge may hold down costs such as employee turnover, while improving performance. JL 

Katia Savchuk reports in Insights by Stanford Business:

Customer agents at a Fortune 500 firm gained a generative AI assistant. The gen AI shared real-time recommendations with operators, suggesting how to respond to customers and supplying links to documents. Compared to workers operating without the tool, those who had the chatbot were 14% more productive. The AI-supported agents handled more chats and were more successful resolving problems. The effect was largest for least experienced workers, who saw productivity gains of 35%. The bot learned what successful agents were doing right. It then disseminated these lessons. With help from AI, agents on the job for 2 months performed as well as agents with 6 months experience. “These are huge numbers. Companies are (usually) happy to get 1% or 2% productivity gains.”

Russia Reinforces Its Bloody Failure In Ukraine's Northeast For No Gain

Russia is attempting to gain a few useless yards of ground to provide photo ops for Putin's predetermined election in March. 

Ukraine is preserving its equipment and the lives of its troops in return for those same battered, smoking heaps of rubble - and thousands of Russian casualties. It's the latter strategy that is a winning one. JL

Nico Lange reports in The Center for European Policy Analysis:

The large-scale Russian attack aimed at the complete capture of Avdiivka has failed. The Kremlin has thereby flunked one of its self-imposed military objectives. Again. Russia has so far lost 20,000 soldiers killed and wounded at Avdiivka, with corpses strewn across the battlefield. 500 main battle tanks, infantry fighting vehicles, self-propelled howitzers, and other vehicles were also destroyed. With this effort, Russia has so far gained 12-15km of land north of Avdiivka without being able to take the city. Ukraine's strategy of not holding every meter, but concentrating on forcing Russia to pay a high human and material price, makes military sense.

Ukrainian Troops' Sushi Obsession Represents Normalcy, Break With Soviet Past

Access to sushi reminds Ukrainian troops - especially those close to the front - and civilians why they value no longer being part of the Russian empire. JL 

Nathan Rott and Claire Harbage report in NPR:

A 30-minute drive from the rubble-strewn hell-scape of Bakhmut, in a brightly lit restaurant on a lightless street, Ukrainian soldiers are waiting for takeout. Sushi rolls. Sixty-four assorted pieces. An artilleryman says "It's important to be able to come back [from the front] and have something from our normal life." The Japanese delicacy rose to popularity in Ukraine following the end of the Soviet Union. Sushi represented Ukraine's efforts to distance itself from its Soviet past. Business is good partly because of the increased military presence in southern Ukraine. And partly because after two years of war, civilians are hungry for a sense of normalcy.

57.1 Percent Of Sentences On Web Incorrectly Translated By AI

The reason for this is that large language models are increasingly being trained with low quality articles which are easier and faster for programmers to use, but which degrade the models' ability to provide accurate, even useful, translations. JL

Jules Roscoe reports in Motherboard:

Researchers at the Amazon Web Services AI lab found over half of the sentences on the web have been translated into two or more languages, often with increasingly worse quality due to poor machine translation (MT), which they said raised “serious concerns” about the training of large language models. A “shocking” amount of the internet is machine-translated garbage, particularly in languages spoken in Africa and the Global South. (But even) in highly multi-way parallel languages, the study found a selection bias toward shorter, “more predictable” sentences of 5-10 words The vast majority came from articles that we characterized as low quality."

____________________________________________________________________________________

A “shocking” amount of the internet is machine-translated garbage, particularly in languages spoken in Africa and the Global South, a new study has found. 

Researchers at the Amazon Web Services AI lab found that over half of the sentences on the web have been translated into two or more languages, often with increasingly worse quality due to poor machine translation (MT), which they said raised “serious concerns” about the training of large language models.

 

“We actually got interested in this topic because several colleagues who work in MT and are native speakers of low resource languages noted that much of the internet in their native language appeared to be MT generated,” Mehak Dhaliwal, a former applied science intern at AWS and current PhD student at the University of California, Santa Barbara, told Motherboard. “So the insight really came from the low-resource language speakers, and we did the study to understand the issue better and see how widespread it was.” 

“With that said, everyone should be cognizant that content they view on the web may have been generated by a machine,” Dhaliwal added.

The study, which was submitted to the pre-print server arXiv last Thursday, generated a corpus of 6.38 billion sentences scraped from the web. It looked at patterns of multi-way parallelism, which describes sets of sentences that are direct translations of one another in three or more languages. It found that most of the internet is translated, as 57.1 percent of the sentences in the corpus were multi-way parallel in at least three languages. 

Like all machine learning efforts, machine translation is impacted by human bias, and skews toward languages spoken in the Western world and the Global North. Because of this, the quality of the translations varies wildly, with “low-resource” languages from places like Africa having insufficient training data to produce accurate text.

 

“In general, we observed that most languages tend to have parallel data in the highest-resource languages,” Dhaliwal told Motherboard in an email. “Sentences are more likely to have translations in French than a low resource language, simply by virtue of there being much more data in French than a low resource language.”

High-resource languages, like English or French, tended to have an average parallelism of 4, meaning that sentences had translational equivalents in three other languages. Low-resource languages, like the African languages Wolof or Xhosa, had an average parallelism of 8.6. Additionally, lower-resource languages tended to have much worse translations. 

“We find that highly multi-way parallel translations are significantly lower quality than 2-way parallel translation,” the researchers state in the paper. “The more languages a sentence has been translated into, the lower quality the translations are, suggesting a higher prevalence of machine translation.”

 

In highly multi-way parallel languages, the study also found a selection bias toward shorter, “more predictable” sentences of between 5-10 words. Because of how short the sentences were, researchers found it difficult to characterize their quality. However, “searching the web for the sentences was enlightening,” the study stated. “The vast majority came from articles that we characterized as low quality, requiring little or no expertise or advance effort to create, on topics like being taken more seriously at work, being careful about your choices, six tips for new boat owners, deciding to be happy, etc.”

The researchers argued that the selection bias toward short sentences from low-quality articles was due to “low quality content (likely produced to generate ad revenue) being translated via MT en masse into many lower resource languages (again likely for the purpose of generating ad revenue). It also suggests that such data originates in English and is translated into other languages.”

 

This means that a large portion of the internet in lower-resource languages is poorly machine-translated, which poses questions for the development of large language models in those languages, the researchers said. 

“Modern AI is enabled by huge amounts of training data, typically several hundred billion tokens to a few trillion tokens,” the study states. “Training at this scale is only possible with web-scraped data. Our findings raise numerous concerns for multilingual model builders: Fluency (especially across sentences) and accuracy are lower for MT data, which could produce less fluent models with more hallucinations, and the selection bias indicates the data may be of lower quality, even before considering MT errors.”