Getting dropped that fast on Realism would have me sick ๐คฆ๐พโโ
MW4 General Feedback - Campaign [October 19]: Enemy AI detection and lethality on higher difficulties could be tuned for fairer, more naturalistic gameplay.
On higher difficulties, especially Veteran and Realism, enemy AI that has been alerted to the player can damage or kill the player unreasonably fast. This includes re-peeking a corner after healing, moving into an alerted room, or even breaking cover at longer ranges. Deaths in these situations happen fast enough that there is little-to-no window to react, which makes enemies feel less like human opponents and more like overtuned bots.
The attached video shows examples of near-instant deaths from my Veteran and Realism runs during the Early Access period (week of October 16).
A blanket increase to enemy reaction time could reduce these unfair deaths, but a few approaches used elsewhere in the FPS genre could keep these difficulties punishing while making deaths feel earned:
1. Accuracy ramp-up: Enemy accuracy starts lower on first sighting or re-acquisition and climbs the longer the player stays exposed. BioShockโs ranged AI did exactly this, starting very inaccurate and tightening over the course of a fight (shoutout @The_One_Boom for referencing this in his recent Campaign feedback video!).
2. Limits on simultaneous attackers: Cap how many enemies actively attack at once, even when a whole room is alerted, so stepping into a more crowded encounter is not an instant death. Two of my favorite series do this: The Last of Us utilizes a Combat Coordination System that assigns enemies to combat roles and caps how many can hold each role at once, and Bungieโs Halo games manage how many enemies are active through task-slot limits.
3. Minimum time between hits: Pace the shots that actually land on the player with a short global timer, so even an alerted room cannot connect several hits on the same frame. Enemies keep firing and suppressing; they simply take turns landing hits. This is a documented approach (see source linked below) with only the enemy holding a โtokenโ able to hit, paced by a timer since the last hit (Game AI Pro 3, Ch. 33). This would not serve to make the player invincible, only ensure a death comes from staying exposed rather than from several hits connecting at once.
The goal is not to make Veteran or Realism easier, only to make deaths feel avoidable, or at least more manageable. Agree or not, I would love to hear your perspective: let me know in the replies how the balance feels to you in the launch version of MW4โs Campaign!
Sources (developers may find these useful!):
-The Last of Us enemy AI design (relevant to #2 above): gamedeveloper.com/design/endโฆ
-Halo 2/3 enemy AI design (relevant to #2 above): gamedeveloper.com/design/comโฆ
-Global time-since-last-hit-timer paired with hit tokens; see pg. 395-97 for information relevant to #3 above: gameaipro.com/GameAIPro3/Gamโฆ
@InfinityWard @CallofDutyCM @CallofDuty
Mikal retweeted
My novella length piece of literary nonfiction will be published soon on here and on my Substack! It's a little below 30,000 words, but in creating it I wrote, and threw away, over 100,000. No part of the piece is written by AI. It's all by my own hand.
This is the most draining creative endeavor I've ever undertaken. The sustained mental exertion over two months of sleeping for one to two hours nightly has finally resulted in something I think is, well, good enough to show the world.
What's the piece about? On its face, it's about a trip to EDC I took with some friends in May 2026. But really, it's about a lot more. More than anything, I want this piece to be interesting for everyone, even those not interested in music festivals. I can only you'll end up as hyperfocused on it as I was.
๐ซถ
Mikal retweeted
PIK 1L financing. We donโt charge you any interest for 3 years but we PIK at 20% with a 2.0x liquidation preference.
I'm taking the workflow side, that's where the money gonna be
An 18-year-old student has created AI-powered robots worth $81,000 that perform their jobs exceptionally well.
Garment factories have resisted automation for decades because fabric is floppy and unpredictable. A humanoid that can handle cloth on machines built for human hands skips the usual answer: a custom robotic line designed around one product.
Picture a factory owner with hundreds of existing stations. Instead of rebuilding the floor, they add robots to the stations already there and shift them between stitching, pressing, folding, and bagging as orders change.
The payoff may not be cheaper T-shirts. It could be shorter runs. Brands face minimum order quantities because changing a line over for 300 units barely pays. If switching garments becomes a software task, a brand could test a design, sell it, and reorder within days instead of betting on thousands of units months ahead. That could even pull production closer to buyers. Nearshore factories rarely win on labor cost, but they could win on speed and lower inventory risk.
A lot still has to be proven: stitch quality, reliability, and who fixes a robot that jams at 2 a.m. Thatโs where a second business appears: technicians, fleet software, and specialists who teach machines new garments.
If the hardware becomes a commodity, the winner may be whoever turns a design file into a packed order fastest.
Would you rather own the robots, or the workflow around them?
Mikal retweeted
Tony Hawk. Bam Margera. One ramp again.
The world is healing.
๐ธ: Getty l @MarkyleRondon
I might actually check out them Alex Xu books, system design stuff got me curious
10 more books I can't recommend enough:
On your way to Software Architect (Part 2)
Here's the order I'd read them in again ๐
๐ญ. ๐๐๐ป๐ฑ๐ฎ๐บ๐ฒ๐ป๐๐ฎ๐น๐ ๐ผ๐ณ ๐ฆ๐ผ๐ณ๐๐๐ฎ๐ฟ๐ฒ ๐๐ฟ๐ฐ๐ต๐ถ๐๐ฒ๐ฐ๐๐๐ฟ๐ฒ (Mark Richards, Neal Ford)
โณ The map of architecture styles and the "-ilities" behind them.
โณ First law: everything in software architecture is a trade-off.
๐ฎ. ๐ฃ๐ฎ๐๐๐ฒ๐ฟ๐ป๐ ๐ผ๐ณ ๐๐ป๐๐ฒ๐ฟ๐ฝ๐ฟ๐ถ๐๐ฒ ๐๐ฝ๐ฝ๐น๐ถ๐ฐ๐ฎ๐๐ถ๐ผ๐ป ๐๐ฟ๐ฐ๐ต๐ถ๐๐ฒ๐ฐ๐๐๐ฟ๐ฒ (Martin Fowler)
โณ The reference for problems you hit in almost every app.
โณ After it, you'll spot these patterns in code you own.
๐ฏ. ๐ ๐ถ๐ฐ๐ฟ๐ผ๐๐ฒ๐ฟ๐๐ถ๐ฐ๐ฒ๐ ๐๐ฒ๐๐ถ๐ด๐ป ๐ฃ๐ฎ๐๐๐ฒ๐ฟ๐ป๐ ๐ถ๐ป .๐ก๐๐ง (Trevoir Williams)
โณ Saga, CQRS, and API Gateway in real .NET code.
โณ The new edition is updated for .NET 10 and C# 14.
๐ฐ. ๐๐ผ๐บ๐ฎ๐ถ๐ป-๐๐ฟ๐ถ๐๐ฒ๐ป ๐ฅ๐ฒ๐ณ๐ฎ๐ฐ๐๐ผ๐ฟ๐ถ๐ป๐ด (Alessandro Colla, Alberto Acerbis)
โณ How to apply DDD to the legacy monolith you already have.
โณ Step by step: from a big ball of mud to modules and microservices.
๐ฑ. ๐ฃ๐ฟ๐ฎ๐ฐ๐๐ถ๐ฐ๐ฎ๐น ๐๐๐ฒ๐ป๐-๐๐ฟ๐ถ๐๐ฒ๐ป ๐ ๐ถ๐ฐ๐ฟ๐ผ๐๐ฒ๐ฟ๐๐ถ๐ฐ๐ฒ๐ ๐๐ฟ๐ฐ๐ต๐ถ๐๐ฒ๐ฐ๐๐๐ฟ๐ฒ (Hugo Filipe Oliveira Rocha)
โณ The real pain of events: ordering, duplicates, eventual consistency.
โณ Shows how to keep data in sync when every service owns its own.
๐ฒ. ๐ฆ๐ผ๐ณ๐๐๐ฎ๐ฟ๐ฒ ๐๐ฟ๐ฐ๐ต๐ถ๐๐ฒ๐ฐ๐๐๐ฟ๐ฒ: ๐ง๐ต๐ฒ ๐๐ฎ๐ฟ๐ฑ ๐ฃ๐ฎ๐ฟ๐๐ (Neal Ford, Mark Richards, Pramod Sadalage, Zhamak Dehghani)
โณ How to split a system and its data when there's no best answer.
โณ Sagas, data ownership, and service size, with clear trade-off tables.
๐ณ. ๐ฆ๐๐๐๐ฒ๐บ ๐๐ฒ๐๐ถ๐ด๐ป ๐๐ป๐๐ฒ๐ฟ๐๐ถ๐ฒ๐, ๐ฉ๐ผ๐น. ๐ญ (Alex Xu)
โณ A 4-step framework to design any system.
โณ Rate limiter, URL shortener, chat. Classic problems, clear diagrams.
๐ด. ๐ฆ๐๐๐๐ฒ๐บ ๐๐ฒ๐๐ถ๐ด๐ป ๐๐ป๐๐ฒ๐ฟ๐๐ถ๐ฒ๐, ๐ฉ๐ผ๐น. ๐ฎ (Alex Xu, Sahn Lam)
โณ Harder cases: payment system, digital wallet, stock exchange.
โณ Worth reading even if you never plan to interview.
๐ต. ๐๐ฟ๐ฐ๐ต๐ถ๐๐ฒ๐ฐ๐๐ถ๐ป๐ด ๐ณ๐ผ๐ฟ ๐ฆ๐ฐ๐ฎ๐น๐ฒ (Lee Atchison)
โณ How to keep high availability as traffic and teams grow.
โณ Teaches you to measure and manage risk, not hope it away.
๐ญ๐ฌ. ๐ฆ๐ถ๐๐ฒ ๐ฅ๐ฒ๐น๐ถ๐ฎ๐ฏ๐ถ๐น๐ถ๐๐ ๐๐ป๐ด๐ถ๐ป๐ฒ๐ฒ๐ฟ๐ถ๐ป๐ด (Betsy Beyer, Chris Jones, Jennifer Petoff, Niall Richard Murphy)
โณ How Google keeps huge systems running: SLOs and error budgets.
โณ Architecture doesn't end at deploy. This book starts there.
๐ ๐ข๐ป๐ฒ ๐๐ถ๐ฝ ๐ณ๐ฟ๐ผ๐บ ๐บ๐ ๐ฒ๐
๐ฝ๐ฒ๐ฟ๐ถ๐ฒ๐ป๐ฐ๐ฒ:
Before you read about a pattern, find where your project needs it.
Books stick when you have a real problem to solve.
Which of these have you read, and which one helped you the most?
He acting like the arcade car just totaled his whole life
DRIVER WILLIAM IS NOT SAFE, HE IS SO LOUD ๐คฃ๐คฃ๐คฃ๐คฃ๐คฃ๐คฃ MEANWHILE HIS FAEN IS SO CALM ๐๐ซณ๐ป๐ซณ๐ป
๐ธ: Aaaahhh! Waaahhh! Help me! Oh my God! How do you brake this thing?! Call the insurance company, please! The car is completely wrecked! Help meee! &@&/@:@..*ยฅ\
๐ฆ*โฆ*
#YouManiacSeriesEP8
Mikal retweeted
she looks like she works at a bookstore and iโm about to ask if they have any books while surrounded by books and holding a book. then buy whatever she recommends. could be an atlas. where are we going.
