Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
TNG5 Profile picture
TNG5
@TNG512
Mar 3 • 7 tweets • 4 min read
@SashoTodorov1 @EmmaMAshford @ConsWahoo The continued rarity of T-72As/Urals in Russian tanks losses strongly suggests that tanks that look bad in satellite pictures really are unusable. So Russia had maybe 2300 usable tanks newer than T-62s in outdoor storage yards in January 2022. @SashoTodorov1 @EmmaMAshford @ConsWahoo "Usable" here having the definition of "can maybe be repaired within 3 years."

Most of the Soviet "modern" tank inheritance was already gone apart from spare parts donor hulls by 2022. More would have surely rotted into uselessness by now anyway.
Jan 25 • 14 tweets • 5 min read
"This is important because if AGI can fully substitute for human labor, then deploying AGI on a large scale will be equivalent to massively increasing the supply of labor, L."

Um, no. AI can only be modeled as an increase in labor if AIs become legal persons. If they aren't, then AIs are capital that makes human labor more productive.

Also, AI that perfectly substitutes for human labor can't exist if only because humans care about and desire social things from other humans, but the problem in this analysis is more fundamental. The key difference is that labor owns things (which makes even human slaves an imperfect analogy given the many historical societies where slaves could own property). So if, for example, we get an increase in labor while an important resource such as land is fixed then the allocation of that resource per human worker will decrease. See IE classical Malthusianism where farms get split up more and more among the heirs with every generation as long as there are more births than deaths.
Nov 5, 2024 • 12 tweets • 3 min read
Or maybe they expected that Zelensky would agree to give up the Donbas well before he agreed to treaty limits on the size of Ukraine's armed forces, and had to readjust after ~week 1 when he agreed to those limits relatively quickly but insisted on a return to the 2/23 lines? Given the recently released draft treaty that demanded that Ukraine just give up all of the Donbas.

I actually find myself now vaguely wondering if the 50k troops limit demand was a bluff, to allow Zelensky to claim victory despite giving up the Donbas and allowing water into Crimea and agreeing to the "reasonable" limits that Putin actually wanted. Maybe limits on missile range and they can't import Western fighter jets, or something.

Then Ukrainian reps unexpectedly offered a 250k troops limit, and Putin thought that such an offer was too good to lose. Problem was, all of their planning for the post-war had assumed that they would be in control of all of the Donbas when the shooting stopped. So they started playing a game of Twister to try to seize the land that they felt they needed in Donbas while maintaining pressure on Kyiv and continuing to hold Crimea's water infrastructure.
Oct 3, 2024 • 7 tweets • 4 min read
NBC News posted this map on February 10, 2022, claiming to be from US intel sources. The "US intel sources" managed to completely miss the attack on Nova Khakovka from Crimea. Furthermore, the text implies that the attack from Crimea was expected to be by air.

"Russian military helicopters would simultaneously support an air assault from Crimea."

nbcnews.com/news/world/u-s…Image They also missed the direction of the attack towards Kyiv, expecting an attack on Zhytomyr.

Which should make one even more skeptical of claims that US intel was able to penetrate the Russian planning process. Far more likely, they based their projections off of the positions of Rusisan military units and so on.
Sep 25, 2024 • 8 tweets • 8 min read
The one major genuine pre-2/24/2022 widespread analytical error that actually involved military capabilities, and not strategic decisions, is that most analysts seemed to assume that the Ukrainians hadn't improved their GBAD that much since 2014.

Here's Justin Bronk writing in February 8, 2022, in which he claims that Ukraine's S-300s (SA-10s) are "largely static due to a lack of key Russian-made spare parts." Most likely this was a genuine problem that they had in 2014 but that they fixed by 2022.

Not universal by any means, Rob Lee for one acknowledged a real possibility that the VKS might not gain air superiority, but sentiments like this seem to have been widespread.

This is an understandable mistake, given that presumably the Ukrainian government did not make their improved air defense systems public knowledge, but I wish that those analysts were more honest about this.

rusi.org/explore-our-re…Image Continuing with Justin Bronk, for example, if he thought that most of Ukraine's S-300s were immobile, then it makes perfect sense why he thought in early March 2022 that "large Russian strike aircraft packages flying at medium or high altitude with escorting fighters would be able to rapidly find and strike any Ukrainian SAMs which unmasked their position by firing at them".

Buk and even shorter-ranged systems wouldn't be able to cover more than a small fraction of Ukraine's airspace, so if Ukraine's longer-ranged systems couldn't move, then yes they would be easy pickings for Russian anti-radiation missiles, and while Russian jets might be in some risk when they came in close to particular high value targets like major cities, which would have Buks and so on as point defenses, then they should have been able to easily dominate the mid-high altitudes over most of the Ukrainian countryside.

So Mr. Bronk's error here was understandable...

rusi.org/explore-our-re…Image
Aug 11, 2024 • 8 tweets • 2 min read
Something that I hadn't noticed in this article before: confirmation that the Pentagon was behind opposition to sending both M1A1 tanks and ATACMS to Ukraine.

newyorker.com/magazine/2023/… Overall, I see little actual evidence for the perception, widespread in pro-UA circles, that Sullivan has been a major brake on US military aid to Ukraine. The article above recounts numerous witnesses to Sullivan pressing more more military aid to Ukraine.
May 28, 2024 • 13 tweets • 3 min read


"INR dissented, arguing that Ukraine would put up a spirited fight and prevent Russia from getting anywhere near the capital. It was right."

No, it was completely wrong. Unless I hallucinated all the video footage of Russian troops in the Kyiv suburbs?vox.com/future-perfect… "Brett Holmgren, INR’s current chief, took pains to tell me that INR was not the only dissenter but confirmed that the bureau thought Ukraine would put up a strong fight."

Nobody actually doubted that. Even the "Kyiv falls in 3 days" scenario estimated 4000 RU KIA.
Aug 28, 2023 • 8 tweets • 3 min read
Current Ukrainian government estimates for monthly Russian missile production: 6 Kinzhals, 30 Iskander-M ballistic missiles, 12 Iskander-K cruise missiles, 20 Kalibrs, and ~40 Kh-101s. They're also refurbishing and upgrading ~10 Kh-22s a month. kyivpost.com/post/21001 In January of this year, they estimated a monthly production of 45 Kh-35s/Kh-59s per month. Assuming zero ramp up, that would be at least 163 missiles produced per month with a range of 200 km or greater. Image
Jul 23, 2023 • 6 tweets • 2 min read
The most recent shooting war between the PRoC and RoC was largely an artillery war over RoC-controlled islands close to the Chinese mainland. en.m.wikipedia.org/wiki/Second_Ta… So leaving aside that as WW2 demonstrated artillery is in fact important in fights over Pacific Islands, and that North Korea still exists, direct protracted artillery wars with the PRoC can't be ruled out either in US defense planning. pwencycl.kgbudge.com/A/r/Artillery.…
Jul 15, 2023 • 15 tweets • 4 min read
Covert Cabal counts nearly 5,000 towed artillery pieces removed from Russian deep storage sites since last February, and 685 self-propelled guns. Seems likely that the towed guns are mostly being disassembled for their barrels, and SPGs to replace losses. https://t.co/nQr7KRZwBv
Image The % of towed artillery among Russian artillery system losses has increased over time. Seems likely that towed systems have been given to mobilized units (early on from DNLR and later from Russia proper) since they're presumably easier to train on.
May 19, 2023 • 6 tweets • 2 min read
Interesting new RUSI report. Among other things, it offers further evidence that older Russian tanks are being used as artillery substitutes, and as assault guns to provide fire support from <2 km behind the frontline. rusi.org/explore-our-re… Image Again, I strongly suspect that Iran was ahead of the curve on this due to their experience during the Iran-Iraq War. Despite producing according to some sources >2000 T-72s from 1992 onward, they keep around large numbers of M60s, T-55s, and T-59s, with domestic upgrades.
May 18, 2023 • 11 tweets • 2 min read
So:

* 5 separate sites visible in satellite photos.
* Shoigu claims that they only fired 6 missiles.
* Russian MoD claims they hit a radar station and 5 launchers.
* 3 large explosions on the ground in videos.
* US claims system is still operational.

Image My current guess on what happened:

Russian intel located the fire control and launcher sites. They fired 3 Iskanders that, instead of warheads contained a large number of decoys released during the terminal phase. Used EW to generate phantom radar tracks.
May 12, 2023 • 4 tweets • 1 min read
175 Iskanders fired in the first month. 829 Iskanders + 123 Onyx missiles fired divided by 175 = 5.44 months. If the total allocation was 850 Iskanders and 200 Onyx missiles, or otherwise added up to 1050, then that divided by 175 = exactly 6 months. 183 Kalibrs+ 86 X-class cruise missiles * 6 months = 1614. 400 Kalibrs + 300 Kh-101s/Kh-555s + 250 Kh-32s/22s + 500 Kh-35s = 1450. I think that they also allocated ~200 Kh-59s, that weren't covered in the graphic. Image
May 11, 2023 • 4 tweets • 2 min read
An article from 2 months ago claimed that the reason why the rumors of Russia buying long-range missiles from Iran never came to anything was due to fears over counter-escalation with regards to Western supplies of long-range munitions. eurasiantimes.com/russia-relucta… Now that it's been publicly confirmed that, if there ever was a deal along those lines, the UK is reneging on it, an obvious thing to do would be to start buying missiles from Iran.
May 9, 2023 • 7 tweets • 2 min read
15 of the missiles were fired at Kyiv, all of which were shot down, so unless there are Western AD systems stationed elsewhere that's another 16 Ukrainian Soviet interceptor missiles gone. I'd conservatively estimate that with the pattern of strikes since April 28, Ukraine is depleting ~0.6 S-300 interceptors per Kalibr/Kh-101 missile Russia fires in this way. Russia has fired 66 since then, works out to 40 S-300 missiles depleted just shooting down cruise missiles
May 7, 2023 • 11 tweets • 2 min read
Turns out that America doesn't have all that many tanks in storage. Binkov estimated current US tank refurb capacity at ~15/month, but would they even be willing to send hundreds of tanks, especially after what we saw with the Bradleys? Image And especially if the imminent Ukrainian counter-offensive isn't spectacularly successful?
Mar 29, 2023 • 23 tweets • 6 min read
After a long Twitter convo with @partizan_oleg, I think I've finally figured out some weird things behind Russia's tank situation.

It basically comes down to a decision made during the Second Chechen War to use a bunch of T-62s for COIN operations. Thread to explain: First up, a thread from May of last year explaining why that decision was made, and how it resulted in T-62s becoming the lowest hanging fruit in Russia's deep tank storage due to them having been reactivated and upgraded relatively recently:
Mar 26, 2023 • 15 tweets • 5 min read
Oryx equipment loss ratios have decreased markedly in all categories except "captured" since September of last year, probably primarily a result of Ukraine using relatively more heavy equipment and Russia using relatively less. github.com/leedrake5/Russ… ImageImageImageImage Total loss ratio shows the same general pattern: Image
Mar 26, 2023 • 6 tweets • 2 min read
Let's do the math on this, using the most generous definition of "produce", so including refurbishing tanks and so on. The Economist estimates that current tank production and repair plants are operating at a surge production of ~950/year. There are probably some diminishing returns as they work on progressively older hulls in worse condition, so let's round down to 900.
Mar 24, 2023 • 17 tweets • 5 min read
Timeline of US military aid to Ukraine.

Prior to the invasion, US government and think tank officials talked about backing a Ukrainian insurgency *even if Russia only occupied part of Ukraine*. March 3, 2022. Russian tanks are in the suburbs of Kyiv. Biden asks Congress for an aid package that includes $3.5 billion in military aid for Ukraine. defensenews.com/congress/2022/…
Mar 23, 2023 • 7 tweets • 2 min read
Since that post was made, we've had additional pledges of:

At least 40 Armoured vehicles
90 APCs
109 IFVs
232 tanks
49 SPGs
Maybe 11 Towed artillery pieces
Maybe 6 air defence batteries Zaluhzny asked for 300 tanks, 500-700 IFVs, and 500 Howitzers. He's going to get more than 300 tanks, but only 289 IFVs and only 237 Howitzers.

I'm a bit surprised. I thought that the US surely has enough mothballed Bradleys, but they might be constrained by crew training times.