Palmz

@RoyalFineArt @ianvisits Is this right? I live right by this and always walk past but didn’t know they had a nice river terrace I could go to!

@ianvisits A similar situation at Devon House in St Katharine Docks, where public access to the river terrace is required (via the lobby) under planning conditions. Always worth exercising the right to stop it withering through disuse.

Exercising my right to enter the Whitehouse Garden behind the Southbank which the owners would rather you didn't, but have to allow it as a planning condition.

@FortKnoxIncome @MerrynSW 40% tax + 9% student + 2% NI, no? I always calculated it as if I was getting 49% of every £1 I earned. Even worse if I’m only getting 43%!!!

@MerrynSW The tax rate for a grad on earnings over £50k is 57% before council tax. It is a genuinely disgusting and how anyone can justify that is bizarre.

80% of grads say wish they had understood loan system better/at all. 85% would vote for a party that did something about their incentive destroying effective tax rates. Great opportunity there...

Looking for 1 ticket entry to Hackney Half - please DM if selling!

@AbdalUllah @TidewayLondon Is there an ETA for completion, and the walkway to Free Trade Wharf opening? Everywhere online still says “done by 2025”

I was honoured to take a sneak preview of King Edward Memorial Park (KEMP) this morning. As part of the KEMP advisory group, we have been working with @TidewayLondon & TH Council for 8 years. few photos of progress. #weloveWapping #weareWapping #LoveWapping

@YIMBYKieran Game of the year contender

I'm unironically so excited

@iifymbro @implode99 @MartonPiller012 Acting like if the mechanic in tarkov was the same in arc people wouldn’t do this exact thing. If you are at extract on tarkov I’m killing them, even with 2 secs in the raid. Welcome to extraction games.

@implode99 @MartonPiller012 nobody does that in tarkov, or at least it's incredibly rare. People extract kill/camp to get your loot, killing someone with 1 second left in raid is just psychopathic and should be universally denounced

They took my babies from me, JUST TO LEAVE THEM ALONE😭

@HK_G36KA4 @Electriz I’m from the UK and bought some - you just need to know someone Chinese. They were all around 50% cheaper then CSfloat

🧵 Some CS2 skins prices based on Buff before/after the update because it just feels UNREAL 📉 Talon Tiger Tooth FN - $ 884 ➡️ $ 198 M9 Doppler P4 FN - $ 2301 ➡️ $ 473 Specialist Gloves Fade FT - $ 447 ➡️ $ 107 Butterfly Doppler P1 FN - $ 4548 ➡️ $ 1193

@nikgeneburn YES

what do you think about this?

@nikgeneburn The survey should have been open from day 1 of the wipe 100%.

ok lets discuss HC wipe here. I'm definitely see that something is good something is not - thats why we testing it. Let us be constructive - those who really want to share how it should be within HC wipe perspective - go ahead.

@BLP_Eng @axel__tv Didn’t play through the wipe as thought the changes were shite and didn’t even know there was an in-game survey, I’m sure 90% were like me who quit early

@axel__tv Probably shit posting anyway but that’s a bit of a leap. Who’s to say the people who didn’t like the wipe and quit didn’t log back on to fill in the survey? It’s all just assumptions.

Glad to see the results, but they are 100% biased and in reality different. Let me explain why. Votes for the survey came only from people that are still actively playing the game, which means all players that tried this wipe and quit within the first 3–15 days didn't vote. Those people did not enjoy the wipe at all and if they did vote, the results would vary DRASTICALLY. As Nikita said, this wipe was a huge success (the number of people playing during wipe day) but how comes only 200K people voted within 3 weeks of this survey being live? (in the past Nikita stated that on big wipes the Concurent playercount hits 400K plus (amount of people playing simaltaneously). To conclude all - this wipe was received very poorly by the playerbase if we account for the bias of the survey.

@YaBoiClear @Jasond68398 @gabefollower 99.9% of people selling a $1000 skin are selling it on a third-party site for actual cash, not for steam credit. Makes way more sense for them to sell direct themselves and get 100%

Do you know how much repeated money they make every time you sell something? A $1000 skin gives valve $100. And the inventory of the marketplace does not change because of it either. So that skin could then sell again for $1000 and steam pocket another hundred. Having an ability to be able to buy a skin flat out with tank the market prices of the skin . Then when somebody would sell it on steam steam would end up getting less overtime because the value of the skin isn’t worth as much.

So apparently this new "Volatile" case allows you to buy one of the skins from the case instead of opening it.

@dosos01964456 @kennycarti1 @nikgeneburn It brought up by Nikita around 2022 as something they were thinking about including, and the wipe video around a year ago kind of teased it

@MP_702 @kennycarti1 @nikgeneburn bruh knocked out 😂😂😂 this aint arena breakout brother

@kennycarti1 @nikgeneburn moving team mates most likely (knocked out mechanic)

@nikgeneburn moving dead bodies so you could loot them safer?

@thegautam Is this deployed and used in the US only? Or EU and/or UK too?

TL;DR: We built a transformer-based payments foundation model. It works. For years, Stripe has been using machine learning models trained on discrete features (BIN, zip, payment method, etc.) to improve our products for users. And these feature-by-feature efforts have worked well: +15% conversion, -30% fraud. But these models have limitations. We have to select (and therefore constrain) the features considered by the model. And each model requires task-specific training: for authorization, for fraud, for disputes, and so on. Given the learning power of generalized transformer architectures, we wondered whether an LLM-style approach could work here. It wasn’t obvious that it would—payments is like language in some ways (structural patterns similar to syntax and semantics, temporally sequential) and extremely unlike language in others (fewer distinct ‘tokens’, contextual sparsity, fewer organizing principles akin to grammatical rules). So we built a payments foundation model—a self-supervised network that learns dense, general-purpose vectors for every transaction, much like a language model embeds words. Trained on tens of billions of transactions, it distills each charge’s key signals into a single, versatile embedding. You can think of the result as a vast distribution of payments in a high-dimensional vector space. The location of each embedding captures rich data, including how different elements relate to each other. Payments that share similarities naturally cluster together: transactions from the same card issuer are positioned closer together, those from the same bank even closer, and those sharing the same email address are nearly identical. These rich embeddings make it significantly easier to spot nuanced, adversarial patterns of transactions; and to build more accurate classifiers based on both the features of an individual payment and its relationship to other payments in the sequence. Take card-testing. Over the past couple of years traditional ML approaches (engineering new features, labeling emerging attack patterns, rapidly retraining our models) have reduced card testing for users on Stripe by 80%. But the most sophisticated card testers hide novel attack patterns in the volumes of the largest companies, so they’re hard to spot with these methods. We built a classifier that ingests sequences of embeddings from the foundation model, and predicts if the traffic slice is under an attack. It leverages transformer architecture to detect subtle patterns across transaction sequences. And it does this all in real time so we can block attacks before they hit businesses. This approach improved our detection rate for card-testing attacks on large users from 59% to 97% overnight. This has an instant impact for our large users. But the real power of the foundation model is that these same embeddings can be applied across other tasks, like disputes or authorizations. Perhaps even more fundamentally, it suggests that payments have semantic meaning. Just like words in a sentence, transactions possess complex sequential dependencies and latent feature interactions that simply can’t be captured by manual feature engineering. Turns out attention was all payments needed!

@thenomadphoto @SecretFlying Same, just checked Oman Air and comes up as Economy but Biz through OTA. Did you speak to the airline/OTA? Looks like I’ll just be getting a refund

@SecretFlying I booked one but it coded as Economy on Oman air but biz on the OTA. Buyer beware!

⚠️ ERROR FARE ⚠️ #BusinessClass from #Rome, Italy to Bangkok, Thailand for only €596 roundtrip #ErrorFare #Travel (lie-flat seats) secretflying.com/posts/error-fa…

@NoTechniqueTV @bstategames How many hours from 0-125 roughly put in?

LVL 125 ✅ The last stretch begins.. @bstategames #tarkovarena

@climate_ben I’m guessing no API for DLR? Great work too!!

This is on the front page of hacker news and coming up on 1mil map tile requests which is getting kinda pricey. Does anyone want to sponsor some tiling credits?

@PalmiraWorld Boat on fire next to the arts centre

What’s today? Helicopter in Woolwich Dockyard. Surely not looking for the girl that drowned in the Thames last week.

@axel__tv Tarkov main could do with a BP, rather than the same quests every wipe as the goal.

Battlepass tease. More to come. imho BP will be released on 7th of April

@absolute_moral @tarkov

@tarkov You shitty incompetents have simply robbed me of 6000 rounds of M855a1, 4000 M61, 4000 300 Blackout AP, 1000 AP SX,.... and much more with your shitty new update, uh stupid losers.

In-game event The Λabyrinth ● The event is available for completion in both PvP and PvE modes. ● Players will receive a unique achievement for completing the event task line. #EscapefromTarkov