SloMoMonday's recent activity
-
Comment on What games have you been playing, and what's your opinion on them? in ~games
-
Comment on Contender for the most non-viable startup: Reflect Orbital in ~tech
SloMoMonday Link ParentSo that was a process optimization technology so not exactly built on any cutting edge research. Rather a way to improve efficiency. This project was for Broiler farms that's used for meat...So that was a process optimization technology so not exactly built on any cutting edge research. Rather a way to improve efficiency.
This project was for Broiler farms that's used for meat production and it is actually slightly less nightmare fuel than Battery egg farms. All farms get their chickens from hatcheries by the truckload (you don't even get them at a spesific qty, they just weigh the trucks on entry and exit and bill you on the difference). Hatcheries supply different kinds of chicken that are selectively bread for purpose. And hatcheries will provide you with a step by step instruction book on the optimal conditions to maximize yield for profitability. That optimal yield involves spesific ratios of meat/fat and muscle tenderness and moisture and a ton of other factors that are used to grade produce so it's not just about stuffing the animals with endless growth hormones. And all this information is in a handy 125page instruction manual. There's also decades of research, manuals, farm records and other datasets with useful information because we've put a lot of work into optimizing food production.
If you look through the manual, you'll see that there are a ton of variables to control, as well as environmental indicators to monitor and adjustments to enact. The most important periods are the first and last 14 days but you can effectively get down to managing every aspect of a chickens life on an hour to hour basis. Many first-world and highly developed farms have been gradually implementing measures and controls to satisfy these conditions and incrementally increasing yields. On the flip side, you have many places where their farming practices haven't changed much since they started operations and there was opportunity to supply them with practical agri-tech that didn't demand a steep capital and time investment.
The start-up wanted to make starter kits of off-the-shelf sensors that farmers could set up in their broiler houses which all connected to a mesh net of hub devices. The entire solution actually lived on these hubs in cases where the farm lost connectivity, but we "needed" everything to go through our subscription service to access the web interface and mobile app. The farmer would map out their broiler house which we virtualized and would dynamically create optimal mounting points for different sensors. Farmer would then allocate the serialized sensors they had available for the BH and we directed their placement for the best coverage. We would then simulate an optimal broiler house with those parameters and flag any significant deltas from the live feeds. The base kit had a ton of functionality and the farmer could supplement it with our more advanced sensors or integrate third party data feeds. And we supported a lot of data points. A lot more than any farmer could possibly need. On top of standard temp/humidity/lighting, we also supported air flow and water flow sensors that was important to minimizing bacteria growth. We could use top down security cams to identify clumping which could suffocate chicks. We even had simple door sensors because letting a draft in can create chaos as well. But I think the most powerful tool was simply providing checklists and reminders at appropriate times since that's the sort of things that are overlooked when managing large operations. The aim was not to increase any individuals workload or add any major new functions to the farm. Just set and forget hardware that would just fire off occasional alerts and it supplemented analytics to identify points that could be improved.
We had working prototypes running across four vastly different broiler houses and there were notable gains over the 6 month trial. I even got a major farm interested and we were ready to test in one of their facilities. But it was all pointless because the start-ups founder was a complete fucking idiot. His grand ambition was to track and trace every individual chicken and have their life story on the blockchain. He pretty much wanted it for the entire agriculture sector. Right from the start his pitch was built around the idea of going to a restaurant, getting a QR code with your meal and being able to track every thing that ever happened to every ingredient on that plate. He also wanted to have a giant "command center" where we would have a view of farms all around the world. My favorite was that we needed to save every single byte of data generated because data was going to be the new currency. All crap that did absolutely nothing to actually help our paying clients.
I can feel my blood pressure spiking just thinking about this guy and all the crap he pulled behind our backs. But it was a pretty interesting project and I actually enjoyed spending weeks on site just talking to farmers and figuring out what was important to the job. I did not enjoy all the techbro startup bullshit I needed to wade through to get to that work and then having all that work be for absolutely nothing.
-
Comment on Contender for the most non-viable startup: Reflect Orbital in ~tech
SloMoMonday LinkI remember this look into the company from a year ago and the company doesn't really make the best arguments for the operation. It's little more than a convoluted and inefficient process to get 10...I remember this look into the company from a year ago and the company doesn't really make the best arguments for the operation. It's little more than a convoluted and inefficient process to get 10 nuclear power plants.
To be most charitable, this could be one of those stepping stone technologies that might lead to something far more practical down the line. Sunlight is technically a signal and there plenty of value across the EM spectrum. Simplist solution would be to ship them off to supplement power generation on the moon or Mars. Or the orbital arrays can be reconfigured for interplanetary data transfers. Or its the building blocks for orbital billboards that will be promising cheap relief to weary space travelers.
It does ultimately feel like wasted resources given the state of... everything and the project seems more concerned with the optics of being an "infinite growth space company". I'd say it's a few steps below OceanGate where it does genuinely have the potential to do some real environmental harm or even be weaponized against critical infrastructure or agriculture. Or as we've seen in Ukraine, it can just be used on the whims of a techbro or dictator without considering the implications on normal people.
Speaking about non-viable startups, that was a dime a dozen in the 2010s. Smart street lights. Smart carpets. Smart tattoos. Crypto banks that are illigal almost everywhere on earth. My favorite was a smart chicken farm system with all the ways to optimize every hour of a chickens life for growth. And it was on the verge of working, just before the founders pivoted to agriculture blockchain technology.
-
Comment on What games have you been playing, and what's your opinion on them? in ~games
SloMoMonday LinkI'd say I'm either about 90% or 40% through Castlevania: Belmonts Curse. I say that because it consistently feels like the game is on the verge of throwing something nuts at you and revealing that...I'd say I'm either about 90% or 40% through Castlevania: Belmonts Curse. I say that because it consistently feels like the game is on the verge of throwing something nuts at you and revealing that everything to this point was tutorial.
I was just chatting to someone about the game and they genuinely only know Castlevania as the Netflix Show. Besides getting to feel my age for a second, it's pretty clear that Konami needs to rehabilitate their IPs.
The Netflix show definitely played a role here and it seems to have set the tone and style template for the series going forward. Every character had a massive chip on their shoulder and theres enough personal baggage in play to fill a large passenger plane. The playable character is also Trevor and Syphas kid so she has a mix of wielding every weapon she gets her hand on while amassing a considerable spellbook through the game.
Theres so many options available that it actually becomes a bit of a pain to manage everything. To the developers credit, there's no shortage of options available between weapons, spells and relics to cater to almost any playstyle. It actually feels a lot like Silksong with the bonus of no limits to when and where you can change your gear. Fully upgrading a spell have little lore drops about how every major characters/bosses got caught up with Dracula. Max level weapons have very powerful abilities and require completing explicit questlines or secret puzzles. But there's no seamless way to use all these things.
You need to go through 3 menu screens for each item. Quick-swap loadouts would have made it so much easier to use the right tool for the job.
Beside that, what seems like very limited enemy veriety and a few boss room run-backs (thank god there's no currency or corpse runs), I really like everything I've seen so far. The entire game takes place in one city but there is a surprising variety of very expansive areas. Each area also has its own gimmicks and the detailed backgrounds make everything seem like environmental storytelling or a clue to a greater mystery.
You also get the satisfaction of solving some of those mysteries before getting the exact clues. A ton of bosses dotted through the map and you have what feels like plenty of flexibility to how you can tackle the main quest. And grinding up to tackle a bad area often involves completing spesific challenges, not running the same loops for exp.
But there's plenty here that feels decisively not like traditional Castlevania and inspired by newer games. Platforming can become brutal and has a very Celeste feel with how precise the movement rules are(there's also a charm that mitigates environmental damage by 75%). Your character also plays a lot more like Hornet, with significant movement tech and detailed animations.
What I enjoy most here is the storytelling. All characters are very emotive and it shows in their dialog, animations and VO. Theres a clear through line of managing ones emotions and the main character is very angry. What seperates the good and bad endings isn't just the hero amassing power, but also learning lessons about how Dracula leveraged powerful people's insecurities. Would have liked to see this play out a bit more in gameplay for the main character, but I guess the whole "easy power" system has been done to death already.
Overall, very fun game. I really hope to see more from the series and for the Dev team to make their own original concepts.
-
Comment on Anthropic logged net loss of over $80 billion in first half of 2026 in ~tech
SloMoMonday Link ParentHere's the direct reporting from Wall Street Journal. The highlight is the massive spike in debt->stock conversions. It's looking to be a big problem given the very impressive revenue numbers from...Here's the direct reporting from Wall Street Journal.
The highlight is the massive spike in debt->stock conversions. It's looking to be a big problem given the very impressive revenue numbers from other legacy tech companies like Alphabet, Amazon and Nvidia. A significant amount of their revenue is a direct result from their stakes in the AI labs and that's distorting a most of their reported figures. Bare in mind, in Q1 alone, this is about $70bil of collective profits (at Alphabet, Amazon and Nvidia) that's not from doing any business (selling ads, chips, licenses, cloud services...), but from the speculation on the unrealized gains. If you had to remove this (speculative) valuation boost and the average earnings are right back down to expected growth projections of 15-16%.
-
Comment on Association for Human Mathematics statement on OpenAI’s October 6 release of mathematical documents in ~science
SloMoMonday (edited )Link ParentMy personal theory is that openAI used a sort of guided-Brute Force method for these proofs and have been sitting on them because they lack the actual explanations for how they got there. OpenAI...My personal theory is that openAI used a sort of guided-Brute Force method for these proofs and have been sitting on them because they lack the actual explanations for how they got there. OpenAI says their model layed out methodical steps to get to the answer, but neglects to share for... reasons? (Theres also a lot of "sorry" in the repo. For context
...sorry, which is a magical term in Lean which provides a proof of anything at all. It is also known as “cheating”. But cheating can help you construct legitimate proofs incrementally: if Lean accepts a proof with sorry’s, the parts of the proof you have written so far have passed Lean’s checks for correctness. All you need to do is replace each sorry with a real proof to complete the task.
I don't have the time to verify how it was used and I should be the last person to judge shoddy math; but I don't think that is allowed when submitting proofs.)
Edit: looked into the sorry statements a bit more and its stated that the "sorry" is to preserve token budget and skip known steps. So it's just going to be extra work to verify, but its not a good look when publishing the papers.
If there was chain of reasoning, that would be front and center of the press release. They would invite the worlds top scientists to do a thesis review with the chatbot. Make it a whole Deep Blue affair. They didn't. We just have a big batch of Kepler Conjectures.
I believe this is the "mokeys with typewriters" idea, but with very smart monkeys who have a good idea on how to write a lean file.
It's impressive in its own right. Likely use one bot cluster of low complexity models to eliminate the 99.99999999% bad possibilities, use others to draft up viable templates and the rest throws everything they have at it until the calculator shows a happy face.
Its not trivial proofs that force the system to return a valid outcome, but its not a useful answer. More importantly, it's not really an LLM at work demonstrating a "deliberate intelligence" which is what they have been trying to sell it as forever.
They also need to save face from the Navier-Stokes situation. This feels like a way of saying "Why would we need to steal this one proof when we have plenty of proofs at home. We just got into the office this morning and there were 400 maths breakthroughs we didn't even ask for."
Even if there was a legitimate proof of a problem, it should be pretty hard for an LLM to "explain" how it got to that truly novel solution because the spesific chain of reasoning wouldn't be explicitly stated in any of its training data. If it was there or there is an obvious causal chain, then it would likely have been discovered by now.
Maybe it could bridge its way to the solution in the way it utilized different concepts to reach the answer, but even that would be unlikely because you need the theory to perfectly match up with the math.
Its important to remember that LLMs don't store and process data. They assess and infer patterns. Very complex patterns arranged in higher dimensional tables and made up of unimaginable quantities of unstructured training data subjected to insane computational power over a long time. But it's still patterns.
1 + 1 = 2 because everyone says so. If the training data is debased to the point where enough of it shows as equal to 15 or or 0 or Strawberry, that's what it will say.
-
Comment on What games have you been playing, and what's your opinion on them? in ~games
SloMoMonday Link ParentI really wanted to get AC8 but that price tag is just a bit steep at the moment and there's seems to be a lack of proper HOTAS support. I did scratch the itch by finally getting to Project Wingman...I really wanted to get AC8 but that price tag is just a bit steep at the moment and there's seems to be a lack of proper HOTAS support.
I did scratch the itch by finally getting to Project Wingman and the AC7 DLC.
Was initially a little disappointed at the early game of PW. The weapon selection lacked veriety. Missions were short, straightforward, had a little too much anti-ground and not engaging enough dog fighting. And theres ob lack of polish like the static briefings, no animations/cutscenes and amateur voice acting.
It's all understandable for a 3 man dev team and what they did deliver was a solid 6/10. Then there's a mission they refer to as The Furball. You break through a cloud and see a mess of aircraft and missle trails and its a complete slugfest. Probably the first time in a jet fighter where I got down to single digit ammo.
Game ramps up from there. Story becomes an earnest MGS fanfic. AOA limiter replaces your characters heart with a high pressure pump. Theres a French woman with a railgun. You get to do a Pearl Harbor and other war crimes.
The only issues is the aforementioned weapon verity and profound lack of 2 Seater Jets (I think it's a deliberate choice to keep players out of overpowered planes) and the devs somehow spilling a bottle of fanta on their keyboard, then it somehow got all over the last few missions.Its just an all out parody of a military with the US (mainly just California) practically taking on the entire world and you get to play the greatest pilot that ever existed.
I think its a good entry into jet fighting games if you want to eventually get to the Ace Combat series. Just avoid the simplified controls and put up with half an hour of crashing into the ocean.
It was a big shock to jump back to Ace Combat 7 because you instantly know that it is a studio effort. Even with dated graphics and lacking some sensibilities from PW (multiple special weps, infinite flairs), you can see budget and experience in the product. Planes have more nuanced controls for each plane and a ton of mods. More complex flight systems (ice, air drag...). More dynamic missions/objectives. It just feels like you're getting a lot more "game" in your game.
In the end, all I did was want to jump into AC8 more. But I'll hold out a little bit. At least until all the DLC and people mod in proper HOTAS support.
-
Comment on What single episode of a show would you recommend as a standalone? in ~tv
SloMoMonday LinkThe Bear S2E6 - Fishes Its technically a prequel episode to everything that happens in the series actually the one I watched first. If you know the show, it's a pressure cooker of tension. If you...The Bear S2E6 - Fishes
Its technically a prequel episode to everything that happens in the series actually the one I watched first. If you know the show, it's a pressure cooker of tension. If you don't it is anxiety inducing family drama.The Bear S2E7 - Forks
And now for something completely different. Seriously, if you are marathon watching the show the flip from the last episode to this one is the biggest emotional whiplash. But it's also an episode that exists amazingly on its own. It's a solo adventure for what is essentially the least likable character in the show. Its almost like a fairy tale mixed with an anime training arc.
On the whole, every solo episode of this series is incredible and makes for a good anthology.
S2E4 Honeydew. S3E6 Napkins. S4E4 Worms.Community S3E4 - Remedial Chaos Theory
At some point Community just becomes a compilation of random standalone stories with the same cast. But this is collection of what-if scenarios that sort of tells you everything you need to know about the characters involved. Also the source episode of one of my favorite memes.Stargate SG1 S4E6 - Window of Opportunity
Its a good time loop episode. I'm a big fan of filler episodes because its the opportunity for character exploration without the constraint of plot. It's something that difficult to do in other mediums and even in modern TV production. And the time loop is just a fun way to see characters give in to their worst impulses and then be forced to escalate just to feel something.Bluey S3E18 - Rain
I'm just a sucker for silent TV and this is an amazing combination of Animation, Music and Storytelling. -
Comment on "What do we do with Jacob Geller?" (Self admitted sexual assault, removed from Nebula streaming) in ~tech
SloMoMonday Link ParentFair assessment and I think the spesific wording here is in reference to thoughts I didn't type. I wasn't trying to defend Geller's actions or imply that the case was perfectly handled as is. That...Fair assessment and I think the spesific wording here is in reference to thoughts I didn't type. I wasn't trying to defend Geller's actions or imply that the case was perfectly handled as is. That should be for the victim to decide.
I should have said that there are abusers that are able to navigate the strict classifications and metrics we use to define abuse in order to effectively maximize the harm they do. By having a broad net to define Sexual Abuse, we loose sight of the systems of abuse that facilitate the dozen rapes that are not reported for the one that is. These systems are hell to dismantle because everyone knows that everyone knows. And in a way, that implicates everyone involved in the abuse done to every victim.
Social media mobs allows people to get the catharsis of sticking it to a single abuser, while allowing the systems and the people that uphold them to make a show of reforming and rehabilitating.
-
Comment on "What do we do with Jacob Geller?" (Self admitted sexual assault, removed from Nebula streaming) in ~tech
SloMoMonday Link ParentI think lot of our societal sensibilities tends to conflate a lot of very complex and nuance topics into a general umbrella of immorality. Harm, abuse, cruelty, prejudice, discrimination. And I do...I think lot of our societal sensibilities tends to conflate a lot of very complex and nuance topics into a general umbrella of immorality. Harm, abuse, cruelty, prejudice, discrimination. And I do believe that a lot of online "discourse" on this topic on either side amounts to performative outrage (this is in reference to knee jerk/meme reactions and opportunistic content creation; not genuine discussion and voicing ones views/experiences. I wouldn't even think of posting my thoughts like this on larger platforms like reddit or even substack).
It's like how the original essay author sort of compares their relationship with his work to the way they see Harry Potter. Works by someone that is deliberately and intentionally dedicating their time and resources to do harm to people that often lack the recourse and platform to sufficiently defend themselves; all while claiming that their cruelty is not only justified, but in service to a greater good. Geller did harm to another person. Assuming even a minor event, it still is a profound and egregious harm capable for permanent damage. A harm that he admits to and as far as we know, one that he is ensuring to never repeat. I'm not saying that the author is wrong to have this view, it's just an example on the gap in proportional nuance to the situation. Both people do harm. Only one I would view as an abuser.
There's also the blatant injustice in the system that plays part in the reaction. Not a lot of people believe that abusers will ever be held to account. Everything about the Epstein and the Trump admin is a testament to that. But also the fact that Geller only received a short suspension. Even if that was enough to reform his behavior and it might have even been a punishment proportional to the harm done. But it also could have been far worse and there are plenty of people that go on to simply resume their behavior after the slap on the wrist. As unproductive as it seems for some people to lean into criticizing Geller, it makes sense when we know mow much worse is treated with the same disinterest.
-
Comment on Anthropic's IPO prospectus shows sweeping AI vision, surging costs in ~finance
SloMoMonday Link ParentI think its just a matter of trust and the most reputable measure I've found is here which reports the figure as: Combined with all the speculation and abstraction around it, plus interesting...I think its just a matter of trust and the most reputable measure I've found is here which reports the figure as:
... taking the last four weeks of application programming interface revenue and multiplying it by 13, and then adding another figure: its monthly recurring chatbot subscription revenue multiplied by 12, according to a person with direct knowledge of Anthropic’s finances. The monthly figure used to calculate recurring subscriptions is based on the number of active subscriptions that day, said the person.
Combined with all the speculation and abstraction around it, plus interesting breakdowns and anecdotes of how people interact with the company; it just becomes really hard to take anything they say at face value. And these are the mass IP thieves who turned a blind eye to AI psychosis and the murder suicides and encouraged the mass layoffs and tortured people with inhumane working conditions and having normal people subsidize their insane utility bills.
Also
Their revenue growing an estimated ~15x in 1 year more than offsets however much their energy costs went up.
Energy costs are only a fraction of the equation. Money supply is the far bigger concern. The raising costs are not just for Anthropic, its everyone and the available free cash for AI investment is going to plummet. It's also going to get very hard to just materialize money like how we have been doing for a long time. Anyone born in the 90's and after has known near 0% interest as the norm and I don't think a lot of new businesses know how operate without that safety net. I think Warren Buffet described it as a Zero G economy. When you turn that gravity back on, we get a good idea of who has brittle bones.
And:
If you think you're unimpressed by fake numbers, VCs are even more unimpressed by fake numbers.
Maybe it's just the part of the Startup space that I dealt with, but in my experience; VC's hate if you're honest about anything. Practically made a point of only considering the prospects that were deemed technically unfeasible or had figures that were blatantly untrue. It's like they had a thing for malicious founders or contractors that would drag them along for years at a time.
I'm genuinely getting worked up thinking about that time. And I know that a lot of my distaste for the AI space is because they feel like the final boss of irrational-tech-startup-cult bullshit. The inevitable conclusion for Blockchain Farming, Crypto "banks" in developing countries and WEB3.0 vaporware.
-
Comment on Anthropic's IPO prospectus shows sweeping AI vision, surging costs in ~finance
SloMoMonday Link ParentAnnual Run Rate feels like a bit of a pointless indicator simply because it's revenue over an unspecified period extended over an entire year. And you can sort of play with it however you like to...Annual Run Rate feels like a bit of a pointless indicator simply because it's revenue over an unspecified period extended over an entire year. And you can sort of play with it however you like to make that number go up. I can say the ARR for my business is $900k because I made $75k in the month of August, but it can also be $2.5mil because I made $7k yesterday alone. I don't see the run rate numbers being based on a spesific time period and its not accounting for any extenuating circumstance in that time. Was it a period of high availability of cheap tokens or was there a new model release or did another company have some bad PR. Why provide an unspecified leading indicator when they can just say the month-to-month revenue over a period or directly compare revenue in a month of this year to the last? Also unlike a lot of other software, revenue is a far less useful metric because there is significant cost to that revenue.
Anthropic is not a stand-in for AI/LLM's as a whole. Investors need to be aware that they are putting money in a company selling a product. A company that incurs costs to manufacture that product plus the costs in providing that product. A company that is subject to global events. A company in direct competition with other companies providing a comparable products.
At some point they will need to generate enough monthly revenue to: Cover monthly vendor/partner costs, service the interest and eat into the principle on all debts, chip away at historic costs to get to that point and have enough left over to stay well ahead of the competition. This does not consider the uncomfortable question of return for investors.Also, the fact that this is last years numbers is very important. Simply because things were a lot better last year. There's the "minor" problem of the Iran war shooting up energy costs. And it's not getting much better with two major waterways at risk and a lot of critical infrastructure either destroyed or at risk and the agri-chem shock only starting to hit farmers and...
But the biggest development between last and this year is interest rates. There's a very big reason trump was so adamant about lowering interest rates. It's because that makes money cheap. And anything is possible when money is cheap (i.e: the startup madness of the 2010's and how all that money pivoted to crypto and streaming/subscription services around 2018). Cheap COVID money is what got the AI industry to this point.
Even though a lot of the AI spend is being reflected in companies cash flows and balance sheets, there is a ton of shadow banking money at play here. Money from private equity and private investment funds; on top of all the related industries and public money tied to it. This is a whole other essay I didn't publish because it snowballed from retail investors getting access to unregulated PC "investments" to leveraged single stock ETFs effectively wiping out a lot of South Koreans savings and started to get into the mass return of Continuation Vehicles which were a centerpiece of 08.It's all essentially a mousetrap reaction. Private credit, AI circular finance, Iran, Energy Crisis, cost-of-living, US national debt, treasury bond bond selloffs... And each of these systems have their own mousetrap reactions inside them as well. For AI there's the Anthropic and Open AI IPOs. There's also data-center roll outs and NVidia chip sales needing to go up into infinity and the looming threat of sweeping regulation and the risk of China flooding the market with dirt cheap compute and dozen other things. Even if AI is the inevitable future of all things, it doesn't change the fact the Anthropic is a very risky investment today.
The meme of line-go-up only goes so far. ARR and monthly active users and global compute demand and total token spend are all cute and impresses VC and press. Investors care about when their money becomes more money. And that's after all the inconvenient things that AI companies don't really want to talk about.
-
Comment on Anthropic's IPO prospectus shows sweeping AI vision, surging costs in ~finance
SloMoMonday LinkOverview of the situation: Reuters got a leaked version of an S-1 document related to Anthropic becoming a public company. Context: An S-1 is a type of "full disclosure" document that states all...- Exemplary
Overview of the situation: Reuters got a leaked version of an S-1 document related to Anthropic becoming a public company.
Context: An S-1 is a type of "full disclosure" document that states all the relevant information regarding a companies financials, revenue, prospects, risks and all the other boring details that are needed to ensure investors know what they are buying. For example here's SpaceX's S-1 and it's... something... (It's definitely not all AI slop, but everything about it felt off.)
AirBnB is a much better example. It's far more coherent filing and they successfully went to market in 2020, in the middle of COVID. I think a lot of the AI labs were hoping to get this sort of response.
My thoughts: Unfortunately, the full document is not available and the reporting on it makes it difficult to really get a handle of what it says about the health of the company. It will be made public a few weeks before the official listing but right now we just have the broad strokes. A lot of headline reporting is focused on the fact that a lot of the document is focused on the EXISTENTIAL RISKS OF AN AI APOCALYPSE. Not exactly surprising given the clickbait economy.
The part I'm personally worried about is the magic math that the company seems to be relying on. Simply put:
The filing offers a rare look inside a business that barely existed several years ago but is growing at breakneck speed due to its role as a key developer of large-language AI models. Anthropic is seeking a valuation of about $2 trillion and has plans to spend hundreds of billions in the coming years to accelerate its growth.
A concept that sits in stark contrast to gems like:
Revenue surged 12-fold in 2025 to nearly $4.6 billion, while operating losses more than doubled to top $8 billion, Reuters reported exclusively on Monday. The US accounted for nearly two-thirds of total sales.
About $3.8 billion in Anthropic's revenue came from customers paying based on how much they use the company's Claude AI system, while subscription revenue came to $789 million. Anthropic said it expects consumption-based revenue to account for "the substantial majority" of its revenue for the foreseeable future.
...
The company paid roughly $351 million back to the platforms in distribution fees, according to a Reuters analysis, suggesting the cloud providers collected some 16 cents for every dollar of those sales. Anthropic reports channel partner fees within the "sales, marketing, and partnerships" operating expense line item on its financial statements.
...
Anthropic said nearly a quarter of its revenue came from two customers last year, and as part of its risk factors, warned that many of its largest clients were not locked into long-term contracts and could cut or stop spending.This is important because a billion dollars is not a trivial number, but the difference between that and a trillion is about a trillion dollars. And leading up to the IPO, the best they can show is 3.8bil in revenue against 8bil in losses.
Everything is happening right as AI Liability becomes a bigger topic, legal/security risks are being taken seriously and global data-center pushback is starting to gain steam. We are also starting to see more demand for models that are cheaper to run, open source and could be hosted locally or on individual machines (if we ever get past the RAM apocalypse).
I think all the AI companies were aiming for the trillion dollar IPO's because for a long time, they represented AI as a broad concept. There wasn't ever an AI narrative, it was just a bunch of vague and sometimes contradictory talking points. AI being able to cure every ailment and make everyone millionaires and get everyone fired and solve the deepest mysteries of the universe and cause vulnerable people to commit suicide and hack every system and solve climate change and killing everyone on the planet. But this is Silicon Valley pie-in-the-sky stuff. Wall Street is the land of bean counters and the beans are not adding up.
The trillion dollar question for them is "how does this technology turn the money spent on it into more money?" It's not just a question about users and customers. There are already a lot of investors and private credit providers and governments and banks who are almost 5 years into this game and without a solid path to profitability and dividends, their only out is the IPO. Some of these funders like the Arab Sovereign Wealth Funds could really use some of that money right now. Same thing with shadow banks and the raising interest rates. And with the whole treasury bond selloff that is not looking pretty right now.
Long story short, all the people who have money in AI companies don't seem to have a way to get anything out of it. These companies are going to public markets and even though the stock market is booming now, it's not going to be an easy ride. Anyone that will be investing in the big AI names now will be taking money out of other places. Places like NVidia, Microsoft, Google, Meta... Yes al these companies have massive AI plays, but they are also mature companies with existing revenue streams. Institutional investors are going to look at these S-1's and ask where is the long term spending commitments these companies have from stable clients? Or what's the cost of training a new frontier model and what's the forecast on seeing a return on that spend? How are 2 super cap companies going to compete with dozens of cheap, small labs that have no trouble stealing their work? Why is Oracle saying that there is a realistic force majeure risk to data-center projects?
Even the regular arguments for AI/LLM feasibility fall short in the face of this massive IPO. Every time token costs go down, the labs will be forced to reduce their prices or be priced out by the competition. The technology getting to the point of autonomous self-improvement still incurs the escalating cost of updating the model and that is still a logarithmic scale, plus the increasing premium for access to high quality training data. As the technology propagates and develops, there are likely to be more companies offering specialized services and it will be increasingly difficult for the major labs to play in every possible space.
The most dangerous part is that many retail investors don't know to analyze a company like this. They tend to work on whatever is in the media hype cycle. Like a lot of the big market action (GameStop, Crypto Boom, Situational Awareness), there's a real chance that regular people get taken advantage of by people who know how to play the market or are victim to the incompetence of people who claim to be in the know.
-
Anthropic's IPO prospectus shows sweeping AI vision, surging costs
30 votes -
Comment on Wikipedia's list of the best video games: how many have you played, and what's missing? in ~games
SloMoMonday LinkI've played most of the list after 2005 and can't argue with almost all of them. Theres plenty of games I would think of adding like Mirrors Edge, Factorio, Power Wash Sim, Celeste, Titanfall 2,...I've played most of the list after 2005 and can't argue with almost all of them. Theres plenty of games I would think of adding like Mirrors Edge, Factorio, Power Wash Sim, Celeste, Titanfall 2, Bastion, Slay the Princess, Rumble VR, Shadow of Mordor, Inscription, Hi-Fi Rush... But it's an endless debate with how many games are out there.
There are a few surprising additions that I love. Valkyria Chronicle is my favorite tactics series and it's a shame that Sonic and Yakuza are the only IPs that exist for SEGA. It's the same with how theres needs to be more of The World Ends with You for people who want Persona without any of the busywork.
Also Streets of Rage was my jam growing up and Absolum is an amazing game by the team that made the revival title. -
Comment on I want an RPG with reverse leveling in ~games
SloMoMonday LinkI think it's difficult to justify stripping away power in an RPG, largely because the point is for enemies to become more demanding in either raw stats or player capabilities. I think the reason...I think it's difficult to justify stripping away power in an RPG, largely because the point is for enemies to become more demanding in either raw stats or player capabilities. I think the reason why so many people feel board of this formula is how it has been over tuned to the point of many game systems being rendered pointless. The "golden path" through an open ended game that keeps you just engaged enough with a steady stream of currency, consumables and gear. A game like Lisa the Painful can end with the main character loosing almost all their limbs but it's barely noticeable in gameplay and combat. Simply because you still power scale at a predictable cadence, regardless of playstyle. I think games like Dark Souls, Fear & Hunger and Look Outside have started to buck that trend and make experiences that always seems to be staying well ahead of where the player is.
A lot of roguelikes also sort of get this right with the "Ascension" system. You keep your meta progression but you can choose to level-up the difficulty in interesting ways. My current obsession is Peglin and the game can get pretty challenging on the base difficulty, with no sure-fire way to stomp the hidden final boss. But I'm on Cruciball 18 and it's a nightmare with your deck being flooded with trash attacks, lack of resources/refreshing and enemies effectively getting 3 free turns when you reshuffle. It demands a lot of high-skill play but the amazing part is that I can still stumble into unexpected synergies and scaling. The problem is with how most experiences don't go beyond a single run so it's hardly an RPG.
Another example is the Warcraft 3 expansion. I think it's the undead campaign where you can watch your commander character be drained of power in real time. Its not utilized too meaningfully since its still easy to just tech-turtle and steamroll every encounter in the short campaign.
Theres also Extraction Shooters where higher level play makes you more of a target and amassing resources changes the way you play just to preserve it. Witchfire is probably the best single-player version of this. I can go into a map to handle some busywork, but as soon as I get a good chunk of currency and materials, the same combat encounters become far more dangerous.
-
Comment on What games have you been playing, and what's your opinion on them? in ~games
SloMoMonday LinkSo I put about 10hrs into the Demo for SlashZero. Seriously, it feels more fleshed out and complete than some fully released games. The best way to discribe it is Side Scrolling, anime, sci-fi...So I put about 10hrs into the Demo for SlashZero. Seriously, it feels more fleshed out and complete than some fully released games.
The best way to discribe it is Side Scrolling, anime, sci-fi Hades (or knock off Dead Cells if that's more your speed). It's pretty obvious after the tutorial and getting into the Home Base area and the subsequent 1st run.
Multiple upgrade trees and hub unlocks for new systems. Different flavors of upgrade that imbues elemental effects on attacks. Set room layouts with rewards telegraphed before selecting. A good stream and veriety of rewards, currencies and powerups to keep you scaling and engaged. Fully voiced and characterized NPCs that can be engaged with gifts and quest lines. A veriety of weapons/classes with unique movement and playstyles (this is through multiple characters).
The selling point is gameplay. It has a style of fluid, animation rich combat that I'd associate with the Dex weapons of a FromSoft game. For example, my favorite character uses a lot of firearms, with a ton of pointless acrobatic flairs and choreography. So a single light attack with a pistol can be seamlessly chained into a follow up and then a delayed medium attack on the third hit. At any point you can interrupt into either a medium shotgun spread shot, or the shot can be charged into a long range rifle shot. But the animations to do that includes a backwards dodge which is not always needed. So to avoid that, you need to do a small jump because the airborne animation is stationary. And then there's a sort of active-reload indicator that lets you chain into quick follow-up shotgun/rifle shots.
All of that is important to keep track of because your character starts of pretty squishy, while enemies tend to hit like trucks, can stunlock you and cram into arenas that border on bullet-hells. You do power-scale quickly enough to overcome this challenge, but the demo only goes to Area-2 ans I hope the difficulty spikes in the full release. I'd also note that the Steam Deck performance is lacking so it's best on PC.
The other game I played was ULTRA0. Was looking for something easy and fun to unwind from a pretty hectic working-weekend. This came up in a random bundle and looked to be a simple "3D" pixel art rail-shooter. Average play time was about 3 hours and I figured it'd keep me occupied on the ride home.
Gameplay is largely forgettable. Zero difficulty. Upgrade system that makes thing trivial within 30min. Largely static levels with a small handful of enemies and music that becomes annoying pretty fast.
I was not prepared for the story though.
Starts off fun enough. You play a cute-high-school-cyborg-orphan-girls that tries to prevent kaiju from breaking containment or else the rest of the world with nuke [unspecified east Asian Island nation] into glass. Chapter 1 starts with your character flight testing a new suit and texting her girlfriend about dinner plans. And ends... in a much darker place.And it's all a roller coaster from there. You jump between performance anxiety and someone trying to save thier idol poster. Grief and substance abuse to a gag about picking up the wrong sketch book. Its the type of anime that would be tagged cute and "psychological" at the same time. Along the lines of Madoka Magica, Shinsekai Yori or Gakkougurashi!.
On the whole, I didn't exactly enjoy my time with the game. I also wouldn't just recommend it to most people and there are plenty of more substantial options for people who are into this style of story. But I don't regret my time spent with the game/story. It does seem like the creators wanted to explore real themes and difficult topics over making pure shock-factor driven trauma-porn.
Its a weird mixed bag. Can't recommend, but it's not something I'd just refund.
-
Comment on Overview of the proofs behind the Navier-Stokes $1 million millennium prize problem in ~science
SloMoMonday Link ParentI don't dispute that most individual inputs tend to get flattened out of general relevance and you're left with an approximate average through the model. But that uniformity is lost with context...- Exemplary
I don't dispute that most individual inputs tend to get flattened out of general relevance and you're left with an approximate average through the model. But that uniformity is lost with context and almost disappears as you get to exceptionally granular topic and niche fields. I'd argue that the spesific input data can only be exaggerated as there are only so many established paths through those topics.
Sort of how you can start by asking an LLM about a story of a boy and get a massive verity of answers. Make it about high school and the outcome window narrows. Set it in the UK, it narrows further and you might start to see some notable trends across output-generations. Take it far enough and eventually you can get to Harry Potter. To be fair there is corrective measures, architectures and formal-verification tools to catch this sort of thing, but the token paths have been (proven to exist)[https://arxiv.org/html/2601.02671v1] in these models training results (and I don't think there's any way around that because these are language-boxes, not analytic engines).
That doesn't even consider the fact that in this case the plagiarism doesn't need to be word for word. We just need to get all the agent logs to verify that the researches works were not ingested and there was no word-for-word reproduction in the reasoning/background processes. Like I said, there are methods to prevent reproduction, but that doesn't mean the research data was not used without consent and for financial gain. I'm actually curious if this whole situation backfires and academics drop the big AI labs entirely with how its looking.
The sequence of events is strange:
OpenAI admits that their work on the Navier-Stokes equation was inspired by rumors that Alpöge and Buckmaster had solved a Millennium Prize problem.
That's immediately a major red flag.
From oAI own report
Nearly 100 agents worked together for approximately 50 hours to produce our Euler regularity disproof.
Once we saw the Euler solution, we thought that Navier–Stokes was the most promising problem to work on. Thus, we decided to devote our resources to Navier–Stokes. To do so, we shifted agents away from the other Millennium Problems and prompted these agents with the Euler resolution. When a further trained version of our internal model became available over the course of the effort, we updated our agents to that model.
So, they had an "internal model" with x-thousand agents running problems. At some point they decided to re-orientate their entire research effort towards a single problem and employed a fresh training dataset. The delta between those training sets is a critical consideration.
According to Buckmaster's statement
My work with Levent has been a purely personal collaboration, free of any
institutional agreements or official involvement by either of our employers. We
used several LLMs throughout: Anthropic’s Claude, OpenAI’s Codex, especially
with GPT-5.6 Sol and, more recently, Astra. The latter was only used for
writeups and auditing our arguments.
...
I asked when the first prompt had been sent by them. This question was
not answered directly by OpenAI for some time. Eventually it was agreed that
it had been sent in the past few days, after information about our work had
reached OpenAI.
I asked whether the model had been trained on, or had access to, our sessions
in Codex, into which we had been putting all our drafts for the whole of this
project. I was told the model did not look up user data. I asked again, about
training, and I did not get an answer.There's also the direct conflict with the previous hype-push of "AI is so smart that it escapes containment and does a bunch of stuff without us knowing". So is that problem fixed and all the AIs pinky promise to be good. We also cant forget the fact that openAI is in a bit of a sticky situation and since revenue is out of the question, the only way for them to IPO is with hype.
And the fact that this is an entire industry built on IP theft. And there the whole issue of how an LLM can be used to generate proofs...
-
Comment on Firefox is releasing an enterprise offering soon in ~tech
SloMoMonday Link ParentMy guess is that there is some anticipation of the Google Antitrust case getting traction again if the Midterms and then 2028 maintains the swing towards more leftist Democrats and the FTC gets...My guess is that there is some anticipation of the Google Antitrust case getting traction again if the Midterms and then 2028 maintains the swing towards more leftist Democrats and the FTC gets enough teeth to put the big tech giants through a wood chipped. And even if that's not the case, it does feel like there are plenty of people looking for alternative browser solutions already.
I remember there was calls for Google to sell off their Browser division and the issue at the time is that there wasn't really a market or competitive buyer for such a product/company (the only options would likely be Facebook, Microsoft, Amazon or the Government; and all of that comes with their own complications). And the retail product doesn't really make sense without the entire advertising panopticon that it facilitates.
There is a good case to make a bid for the enterprise portion alone. Primarily because its the type of boring, background, mass scale, 5+ year b2b licenses that make for successful software companies. With some enterprises starting to sour on the perpetually escalating SaaS pricing and having critical functions subject to the whims of host companies, I can see a market for a good on-prem admin suite (with a comparable and siloed cloud option for SMMEs). Bundle in a strong pedigree, out-the-box GDPR/CCPA compliance, Ad blocking, anti-AI tools, VPN, provider agnostic email clients, open source office intergrations and all the other things Google can't exactly include in their own product; it becomes a feasible option for anyone looking to untangle from Single-service-provider dependencies.
And enterprise is also a very good gateway into the retail market. If a user can seamlessly create a work instance through their personal-device client or if Firefox enterprise becomes the standard browser at schools/training, it just nudges more people into the ecosystem. Hell, I know people will migrate the second they see an ad-blocked internet on their work computers and learning they can get the same anywhere.
-
Comment on What made that wedding that you attended particularly memorable? in ~life
SloMoMonday LinkWhen my dad remarried, my siblings and I spent the entire event telling everyone that it sucked when we were not invited to the first one. Since we are all adults now, it's pretty hard for people...When my dad remarried, my siblings and I spent the entire event telling everyone that it sucked when we were not invited to the first one.
Since we are all adults now, it's pretty hard for people to imagine us as not being born at the time.
So responses ranged from, "Ha ha. Good one.", to disinterested eye rolls, to my uncle having a little crisis trying to figure out how we could possibly not be invited to our own parents wedding.
I'm about 15hrs in and no where close to the tower.
Granted, I have a habit of sequence breaking and abusing geometry to get places I really have no business being. Its also pretty clear that there are systems just begging to be broken and you end up in interesting places. I left the fist proper ocean-dungeon for a few rounds of herb farming and decided to try my luck with the obvious suicide path. Next thing I know, I've wrong-warped to the end of the game and I'm sitting with a 6% achievement.
The best part is that it wasn't a bug or accident. From a design perspective, theres so many little things that feel "off" about how various functions are executed. It doesn't take a lot to figure out that it's all deliberately set up for some Outer Wilds shenanigans. Similar thing with how important item discreptions and environmental clues are.
A little note on some mechanics I figured out: the Endu sword actually is a very powerful dodge item. Unlike most other weapons, it's secondary attack will push you in whatever direction you are moving in, so long as it has charges. So you can lunge in any direction and its important for the more intense fights. But it doesn't work when you air-dash with the dedicated key.
Also, the air dash will push you downward if you even glance in that direction. Look up when trying to traverse horizontally.
And if you are on the back foot and retreating, frisbey throwing weapons evens the playing field.
Lastly, Duck Merchant is the best merchant.