April flowers did bring May flowers, after all. Speaking of, this piece is brought to you by August Fanon’s The Golden Child — Nas [FANON EDITION] Vol. 1: a lush, lavish, and glossy mixtape by the Brooklyn-based hip-hop producer reimagining some of Nasir Jones’s most significant tunes off Illmatic (1994), It Was Written (1996), and The Lost Tapes (2002). Should you have twenty bucks to spare, do your daily good deed and dash them toward your main man August by copping his tape instead of funneling them into prediction or crypto markets. He’s fighting the good fight for us.

At any rate, fights rule everything around us, and if you’re anything like us you’ll know of some highfaluting ones of the sporting affectation taking place right now as the National Basketball Association entered its momentous 2026 post-season phase a few weeks ago. Last month, we wrote and didn’t regret our bold NBA Pick ‘Em Bracket Challenge predictions, handily nailing the left-side Western Conference half of the Merkle tree, yet completely botching the East through what at the time didn’t seem like such outlandish picks by placing Detroit and Boston in the Conference finals (the first and second regular season seeds, respectively…?). Nope, both teams clearly had other plans and prioritized getting a head start vacation sign off to Cancún, with the Pistons imploding on Cunningham’s countless turnovers and Durens complete disappearance off the face of the earth, while the Celtics self-sabotaged via a mixture of prepotence, in-fighting, and folding Tatum too early back into the team.

We like to think we’re at least partially redeeming ourselves by having a 100% Western Conference finals hit rate with the somewhat predictable landing of both OKC and the Spurs (the first and second regular season seeds, respectively…!), as well as an active NBA Champions pick still in the race in Wembanyama‘s jackals squad. Sticking to the left-hand side of the Association (no relation to political affiliations, NB), we’d even have a perfect bracket all the way to the first round match ups, had Kevin Durant not willingly chosen to check out for the year before the Rockets series against the Lakers even began. Like, Los Angeles actually advanced 4-2 basically without Dončić and Reaves, facing little to no opposition by an unrecognizable Houston team, only to get wholeheartedly obliterated by the reigning Champs Thunder in the next round. Did we even mention we’re New York Knicks fans over here? We know you wouldn’t tell looking at our predictions. That’s why we keep off Polymarket.

Amidst this bulletin, we did see fit to return to our only friend, the NBA AI chatbot—and us likely being its only friend too, considering the radio silence around it this whole season (the service still sports a ‘Beta’ label next to it). So we went ahead and asked it some more questions. Only this time sprucing some Playoff zhuzh on top. We broke the queries down in three packages; two questions for each of the Conference final match ups, as well as two bonus missives we thought would throw it a bit for a loop. As per usual, you’ll find this mysteriously underreported feature by going to the Discover tab in the official native NBA App, and selecting the apposite card from under the Around the League menu. All questions and answers are up to date as of 21st May, with stat lines and data allegedly corresponding to the same timestamp. We started with our beloved East, looking for a hybrid between a low-key scouting report for each team and gratis intel to each opponent’s coaching staff:

First of all, this thing seems to have improved. Not only are we not sure this type of prompt could even have been processed and answered by the LLM five months ago, but some of the responses returned by our bot friend actually held some water. Let’s pick out some highlights. According to the NBA (?), in order to bounce back from the historical remuntada suffered at the hands of New York a couple days ago, Kenny Atkinson and his Cavaliers should i) attack Brunson’s lackluster defense more deliberately, ii) actually deploy healthy players they have on the bench such as Tyson and Ellis, iii) maximize the floor and be more unforgiving vis-a-vis the Knicks‘s tag teaming, iv) place former Defensive Player of the Year Evan Mobley under the rim, and v) freaking do something about James Harden. Aside from the pretty sus underplaying of the Madison Square Harden problem on account of how blatant it was, these are legitimately valid improvement areas for Cleveland—if slightly basic.

Hilariously if you’re the Cavs, the list of AI recommendations for the Knicks to keep the momentum going and leave Gotham City up 2-0 on the series start loudly and proudly with a “[k]eep attacking James Harden in pick-and-rolls”. While Mike Brown’s squad did receive an extra solicitation bullet point compared to their rivals (six to five), it did come in the form a trippy contradiction between its ethos (“Force Cleveland’s bigs to protect the rim and rebound in space”) and the preceding advice (KAT pulling Allen and Mobley away from the rim). Pick one lane, botty mouth. The final three recs all seemed based though, focusing on New York’s rapid, visceral, and dynamic defense, their sharp three-point shooters, as well as resorting to Brunson’s hero ball late in the game. Again, pretty duh takes—not least considering the Knicks captain was named Clutch Player of the Year last year—yet not too terrible considering this is free advice. Let us hop over to the West next.

Unlike the skewed critical mass noticed for the Eastern Conference finalists, both San Antonio and Oklahoma got six notes each, as they just rebalanced their series score after the instant classic game 1 which saw Wemby and co. edge out the reigning champs. Because the French alien is the best player on the planet right now, we chose to word the prompt leading the OKC’s scouting report around potential ways to stop him, rather than the more generic formulation reserved to the other teams. Yup, his gravity and impact is that strong.

This second batch of responses came through as more of a mixed bag. Firstly, while it’s somewhat common knowledge that Gregg Popovich still acts as the Spurs whisperer—surely commands a larger-than-life presence in the organization—for the Association’s official AI chatbot to still bill him as the head coach in two different passages is flat out gross and unforgivable. Not only is it disrespectful toward Mitch Johnson’s excellent work since taking over the honcho position last year, but it seriously throws into question the kind of data this language model is using and training on… Anywho, notwithstanding such faux-pas, the pieces of advice to San Antonio are generally fair game and strike a cool balance between offensive (protecting the ball, avoiding turnovers, spreading the floor in half court sets) and defensive (return to dominate rebounds, contain pick-and-rolls for SGA) adjustments.

As far as neutralizing Wembanyama is concerned, according to the chatbot Oklahoma should… throw more physicality at him? Uhm, firstly; did Daigneault not think about it already? And secondly, how exactly? Not even OKC’s larger-than-average bigs Holmgren and Hartenstein seem to have figured out how to slow him down—they got a game each to attempt that, unsuccessfully so. If anything, the last two bullet points appear a tad more conducive to scalable in-game adjustments SGA and friends could leverage depending on how brutally Wemby annihilates them: a) picking up pace offensively, beating him and the Spurs to the spot, and b) exploiting switches and screens to minimize his presence and gravity around he basket. All easier said than done.

Lastly, we did tease two bonus prompts for each of the two Conference finals, and this is how we put NBA’s AI to work: when asked about the potential impact of the Western finals’ IL—including projected starters De’Aaron Fox, Dylan Harper, and Jalen Williams as of the time of this writing—we were left pretty underwhelmed by the microwaved reflections yielding neither fish nor fowl (“those three injuries leave both teams short-handed”; “Spurs struggling to organize their offense”; “the Thunder needing others to replicate Williams’ defensive and slashing presence”). You know that feeling you get nowadays when AI spits back those hollow, boilerplate, surface-level, yet elegantly written (and perhaps “believable”) answers? These bonus prompt felt the most like that.

Even the more interesting and speculative inquiry about who the Knicks—billed by everyone as the clear East favorites to reach the NBA Finals for the first time in almost twenty years—should prefer to match up against in the prize fight was not only safe and uninspiring, but riddled with factual errors (wrong Playoff pace stats compared to OKC and average turnovers per game) as well as omissions (perhaps mentioning the fact that New York already beat San Antonio in a high-stakes finals game this season would have been appropriate?). Oh, lest we didn’t make it clear, AI too thought it was the Spurs.

So where does this leave us exactly, five months removed from our first sandbox trial of this little known, poorly-advertised proprietary algorithmic feature of the league? We’re not sure, actually. The bot’s answers did appear to become more substantial and articulate, at least at face value. Unlike our first underwhelming date, the NBA’s LLM didn’t push back on any prompt this time, no matter how forward-looking or potentially incriminating they might sound be by being associated with the Jerry West logo (we even asked it for its thoughts on who will lift the Larry O’Brien Championship Trophy this year, and it docilely returned bookmakers’ odds with the usual blah blah blah—not pictured here for obvious reasons). Yet, we’re still not sure we came away with learning anything new, or different, that we couldn’t just as easily have scraped from the right corners of NBA X (or Bluesky! Yes, yay, Bluesky is rife with hoopin’ blabber these days, believe it or not!). This features remains a modern mystery to us.

All things considered, does this piece double as an indictment of LLMs? Well, did y’all see the gnarly AI-generate companion image below? This tech can’t even spell the word Oklahoma correctly—how can we trust it with deep Playoff on-court lore and intel?

We’d like to thank you sincerely for taking the time to read this and we hope to feel your interest again next time. And let’s go Knicks, this time around.

AV