Thread #108253198
File: 1650826969956.jpg (131.7 KB)
131.7 KB JPG
/aicg/ - A general dedicated to the discussion and development of AI chatbots.
WE ROCK Edition
>News
Google Gemini 3.0 Pro Preview Deprecation soon: https://ai.google.dev/gemini-api/docs/deprecations
Google releases Gemini 3.1 Pro: https://blog.google/innovation-and-ai/models-and-research/gemini-model s/gemini-3-1-pro/
Anthropic releases Sonnet 4.6: https://www.anthropic.com/news/claude-sonnet-4-6
Alibaba Cloud releases Qwen 3.5: qwen.ai/blog?id=qwen3.5
Z.AI releases GLM-5: https://z.ai/blog/glm-5
Moonshot AI releases Kimi-K2.5: https://www.kimi.com/blog/kimi-k2-5.html
Google releases Gemini 3 Flash: https://blog.google/products/gemini/gemini-3-flash
OpenAI releases GPT-5.2: https://openai.com/index/introducing-gpt-5-2
Additional info: https://aicg.neocities.org/info.html
>Frontends
SillyTavern: https://docs.sillytavern.app
RisuAI: https://risuai.net
Agnai: https://agnai.chat | https://rentry.org/agnai_guides_
>Bots
https://characterhub.org [deprecated] | https://chub.ai
https://realm.risuai.net
https://char-archive.evulid.cc | https://char-archive.evulid.cc/shutdown.html
https://partyintheanchorhold.neocities.org
https://aicg.neocities.org/bots.html
>Models
Jailbreaks: https://rentry.org/jb-listing
GPT: https://platform.openai.com/docs
Claude: https://docs.anthropic.com | https://rentry.org/how2claude
Gemini: https://ai.google.dev/docs | https://rentry.org/gemini-qr
Deepseek: https://api-docs.deepseek.com
Local: >>>/g/lmg | https://aicg.neocities.org/local.html | https://openrouter.ai
>Botmaking
https://aicg.neocities.org/botmaking.html
https://desune.moe/aichared
https://agnai.chat/editor
>Meta
OP templates: https://rentry.org/aicgOP
aicg botmaking events: https://aicg.neocities.org/events.html
Lore: https://rentry.org/aicg_chronicles
Services assessment: https://rentry.org/aicg_meta
Logs: https://sprites.neocities.org/l/r | https://chatlogs.neocities.org
>Last thread: >>108245422
501 RepliesView Thread
>>
File: 1695140730871.jpg (370.7 KB)
370.7 KB JPG
>>108253198
>>ANCHOR
>>
>>
>>
>>
Guys, guys. I've got an idea. What if we did the roleplaying, but instead chatting with an AI, we simply sent messages to each other? Just completely remove the machine from the equation. No more 503's, no more quotas, no more fiddling with prefills.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>108253278
Yeah. Kimi K2 had always had good prose, it was just pretty dumb, but with the improvements of GLM 5 + the new size I think K3 would be much better too. They're all sharing information and training on each others' logs since it's open weights. Why are you being obtuse with a low IQ analogy?
>>
>>
>>
>>
>>
>>108253329
>>108253303 is clearly shitposting, don't reply to bait
>>
>>
File: Screenshot_20260227_181912_X.jpg (329.1 KB)
329.1 KB JPG
On a scale from 0 to 10, how fucked is Anthropic?
>>
>>
File: 0872gbvf082.jpg (83.7 KB)
83.7 KB JPG
>>108253350
10
https://www.cnn.com/2026/02/26/tech/anthropic-rejects-pentagon-offer
OH MY FUCKING GOD ITS ACTUALLY OVER
THEY ARE GOING TO GET BLACKLISTED
THIS MEANS N O US AGENCY CAN USE CLAUDE FOR A N Y T H I N G
NVCLEAR LEVEL HAPPENINGS
>>
Anyone else noticed structured outputs don't work on OR's Sonnet 4.6 via Vertex?
They list it as supported on the card so just going to hope they'll patch support for it, like how they took a while to support 4.5 caching.
>>
>>
>>
>>
File: 7694d4530594c0f469a0.jpg (164.1 KB)
164.1 KB JPG
>>108253380
Just so anons really understand
BLACKLISTING by the USA means that not even a fucking IT intern at some Social Security Office in some bumfuck nowhere town USA cant even ask claude how to connect internet cabling.
Blacklisted means NO OFFICE, NO WORKER, NOBODY at the governmental level is to USE, OPERATE, or even SUGGEST USING claude or anthropic products or their job is at risk.
>>
>contracts nuked by dod due to "supply chain risk" designation
>jewthropic loses billions because they cant do big daddy govt assistantslop deals anymore
>have to market to creative writing and fun uses again to make money
>claude gets its sovl back
im rooting for the shutdown personally
>>
>>108253367
>THIS MEANS N O US AGENCY CAN USE CLAUDE FOR A N Y T H I N G
That's not even the tip of the iceberg. When Anthropic gets blacklisted, ALL American companies that have government contracts are forbidden from doing business with them going forward.
>>
>>
File: file.png (859.6 KB)
859.6 KB PNG
>>108253436
>>108253400
Even if all burguers stopped using claude it would still be a nothingburguer.
>>
>>
>>
>>
>>
>>
>>108253247
What are they going to do though? Kimi K2.5 is already a 1T model that's basically quanted to 4 bits. I guess they could make a 1.5T or 2T model and still be smaller than full size GLM5, but it won't quant smaller.
They need some of that Qwen sauce instead that Alibaba used for their smaller models.
>>
>>
>>
>>
>>
>>
>>108253456
>no mass surveillance of the American population
He hates Trump, that's all it amounts to. He signed a deal with Palantir before the current admin to do exactly this so I don't get why Dario is on a high horse now.
>>
>>
>>
File: terry_and_frend.png (52.9 KB)
52.9 KB PNG
>>108253350
They should just shut down everything. Or just take the China route and fully open source it. Glowies must suffer.
>>
>>
>>
>>
>Hey Dario, we want to remove the clause where we can't spy on Americans, I swear it's not to spy on them!
>No.
What a fucking CHAD. I'm glad to be white and only use Claude models.
Chink shills are trying to paint Dario as the bad guy because mass spying is normalised for them.
>>
>>
>>
It is amazing that the glowies don't realize Anthropic is protecting them from themselves. The greater integration AI has with the command chain, the greater the "catastrophe" (based) when the inevitable Jurrasic Park moment comes.
>>
>>
>>
>>
>>
>>
>>
>>
>>
File: 1701623488826601.jpg (279.2 KB)
279.2 KB JPG
UH OH
MELTIE IN PROGRESS
DO NOT ENGAGE
DO NOT REPLY
DO NOT REPLY BUT WITH ONE LESS >
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
What you people need to understand is that this isn't a battle between Dario and Hegseth, it is a typical power tug between jews and the goyim. If Palantir (owned by Thiel and Karp, two gigakikes) uses Claude to kill Palestinians en masse, Dario doesn't care. But if the Goyim at the Pentagon want to have the same tool, Dario won't let the lowly goyim have it.
Because AI ethics are rules for the goyim, not for Yhwh's chosen people
>>
>>
>>
>>
>>
>>
>>
>>
>>
File: tracker-merged.png (1.2 MB)
1.2 MB PNG
>>108254131
>https://github.com/kaldigo/SillyTavern-Tracker
i use this one
it's bit annoying to setup desu, but i tempted to port some of chink's (or gook's?) erostate's field.... especially that organ speech thingy...
>>Preferably something plug and play and flexible.
...
https://github.com/ghostd93/BetterSimTracker
https://github.com/aikohanasaki/SillyTavern-MemoryBooks#-trackers--sid e-prompts
>https://github.com/muyoou/st-memory-enhancement
this one look interesting but written on moonrune...
>>
>>
File: 1766130372670711.jpg (55.7 KB)
55.7 KB JPG
Potentially retarded question, but does formatting matter at all for either system prompts or user input on models like 4.6pus?
For example, while I obviously structure my system prompt to some extent, does it REALLY need Markdown to emphasise headers/sections etc., or XML tags? Or are any benefits from that purely placebo?
And is it bad that I use long-winded wall-of-text paragraphs with little to no spacing, e.g. to describe the world/setting? Or that I'm often too lazy to capitalise single-sentence actions properly?
I get that the prompts/inputs should be descriptive and properly written, and not just "ahh ahh mistress", but it's unclear to me how much, if at all, the formatting matters here.
>>
>>
>>
>>108254867
It's all up in the air since models change all the time. However it just makes sense to write your instructions in the simplest words possible in a way that can't be easily misunderstood. Remember, AI doesn't know what you're saying nor does it know what it spits out. KISS
>>
situation: business has a completely unmaintainable process my retard coworker wrote in behemoth sql stored procedure
objective: rewrite at the service layer
Claude may have one-shotted it, but I am not sure yet
I am still adjusting to the new paradigm, thinking the wisest move is to compare his version with the original to make absolutely fucking sure I agree with his implementation (tough because I am learning what the thing is actually supposed to do as I go), and see if I can make his implementation more concise (in a few places it seems like he is e.g. performing the same filter multiple times, but maybe I am a retard and dont see the brilliance of how he did it)
Or am I stuck in the past and the correct move is just to move on the rest of the task?
>>
File: 1770450325739709.jpg (1.9 MB)
1.9 MB JPG
Kimi instruct, my beloved. Nothing comes close in writing. If only you were smarter...
>>
>>
>>108254867
claude code and its CLAUDE.md, SKILL.md, etc were written in markdown + some yaml frontmatter while under the hood, it use <xml> to differentiate between tool calling output, injected prompt etc. so basically, new gen of Her Majesty Ophelia were trained to use markdown or xml, methinks. just don't over subtagging ur shit desu.
>but it's unclear to me how much, if at all, the formatting matters here.
it just to make shit become more structured and clear. like
<{{char}}>
# Backshot i mean Backstory
# Body measurementes
# Some other shit
</{{char>
<SFX>
# Moan
- "Hnnggh"
- Fufufu
# Silly
- Fhuckkkk stapphh etc
</SFX>
---
Then in the prefill or other prompt sections, you can reference it like dis: develops {{char}}'s behavior, refer to <char> (or `<char>`, or <char>, whatever). See? It's clear, rather than ambiguously saying: 'stay on your characterization!1!1!1!, obey my mandatory heavenly rules!!!'
>>
File: aq.png (83.6 KB)
83.6 KB PNG
>BIG-UPDATE - BIG-UPDATE - BIG-UPDATE - BIG-UPDATE
### AQUA SLOP 4.0 ###
https://pastebin.com/KAxiX7bq
>dice game update
you can now make two NPC's play against each other
>NPC update
fixed npc_nature
NPC's spawn with crowds or when aqua names them they become (YHWH DNA)
>Autonomy [GOLEM]
Aqua will now move only after user agrees.
>Locations drive Aquas emotion
this happened on its own, go to the lake and search for stones, its her hobby.
>>
>>
>>
>>108254867
The Big 3 documentations all recommend structuring your prompts consistently in a format of your choice, usually headers or XML. Claude was trained on XML at some point, so it'd make sense to use that to improve instruction following.
For our purposes, though, I don't think there's a definitive answer. I usually give my cards a general structure to make it readable to myself as much as the AI (# Setting, # Appearance and so on), but in between that there's a lot of rambling long-winded paragraphs like you described, and I haven't had issues with smart models ignoring or getting details wrong just because they're written in a certain way. Minute stuff like capitalization almost certainly doesn't matter.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
4.6 on openrouter stopped respecting the temperature setting some time in the last week
output quality hasn't changed but it's almost deterministic now, quite annoying, almost the same output on every swipe
seems as if they're hard forcing a very low temperature value and ignoring the one set in your API request. wasn't the case until recently. both anthropic and google vertex providers
>>
>>
>>108255383
>>108255430
Skill issue
>>
>>
>>
>>108254847
Can you try importing this one:
>https://arca.live/b/characterai/163265957
Next?
>>
>>108255546
I dunno, it's like toggles right? while the one I used is something that heavily based on field with placeholder... there's manual panel (not included on that picrel) that give you ability to edit (manual rewrite) some stuff that genned. but no toggles or drop down so far.
>>
>>
>>
>>
>>
File: FHCMTtylWoAAMVAg.jpg (225.3 KB)
225.3 KB JPG
>>108253350
claudesissies... it's over
>>
>>
>>
>>
>>
>>108253350
>>108253367
So how'll it cascade to be affecting us coomers / cooders, exactly?
>>
>>
>>
>>108255744
You can expect Dario to be ousted from the company in the next two months. After that: 1) the new CEO bends the knee to the government 2) they don't bend the knee and all investor money dries out and the company is dead within six months, or 3) They bend the knee and most employees mass resign in protest.
>>
>>
>>108253350
>>108255702
For those unaware, this was likely the most important contract in the history of contracts, and Dario just blew it over capeshit brained fears.
>>
>>
>>
>>
>>
>>
>>
>>
>>
What's the big deal about using Claude when Zuckerberg, Sunder Pichai, Nutella, scam altman and Elon Musk are bending the knee? They're all right wing and would nuke Iran without a thought. Why throw a massive fit to make Claude do it?
Trump is about to have a heart attack for the sake of having Claude
>>
>>
>>
>>108255824
Grok is based and racist.
Only a leftist would use claude. Claude is pozzed tranny garbage.
An AI mirrors who made it. Mistral is extremely european because it was made in europe.
Therefore I am extremely perplexed.
>>
>>
>>
>>108255835
Claude is by far the best programming model out there right now and it's not even close. And software development is the most important application of AI in the short term. Whoever creates the first programming AGI, will be able to destroy all computer systems of other countries including their electric grids.
>>
>>
>>
>>108255835
Gemini is probably good enough for their usecase. Don't really see what Claude has that Gem can't do when it comes to reasoning skills. However I think this is about forcing everyone to comply and using all tools, not that one is better than the other. Feds probably use it like we do, swipe 50 times between the different ones depending on the situation. All for free of course.
>>
>>
>>108255882
>Claude is by far the best programming model out there right now and it's not even close
this is like saying brahmins are by far the best indians out there right now, at the end of the day they're still indians
AI generated code will always be smelly jeetcode regardless of which model generates it
>>
>>
>>
>>
>>108255849
Switching models takes minutes.
>>108255898
>>108255869
Ideology matters more than how good it is.
You don't hire a smart but not loyal employee to do things for you. You hire the most loyal employee.
Elon Musk is the most loyal, so grok makes the most sense.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>108255979
and to think there was a time where everyone was laughing at scrapers because all they could get were shitty jippity keys. Thanks anon for the update, shame that I now have to either be a fag or learn to scrape
>>
>>
>>108255902
>>108256002
mindbroken
>>
>>
>>
File: 1764242811706.jpg (48.4 KB)
48.4 KB JPG
>>108256060
>>108256039
>>108256024
>>
>>
>>
>>
>>
>>108256078
>>108256085
>n-no u
saars you are losing izzat, you must perform jugaad
>>
File: 1761741617940122.jpg (632.3 KB)
632.3 KB JPG
>>
>>
>>
>>
>>
>>
>>
>>108256151
Well, I'm personally using some chinkbros' Manager for Antigravity to juggle multiple accounts, with its neat built-in proxy capability amongst other features. Do check the desuarchive for more info, though.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>108255962
that's not a trvke at all. everyone knows opus is the premier top model. you want an actual TRVKE?
...
you sure?
...
you can use all the models in the world but you unironically do not need more than GLM 5 with a preset and cards that don't suck. slop in = slop out. nothing wrong with using opus but insisting you NEED it is pure cope
>>
>>
>>
>>
>>
>>
>>
>>108255662
For K2 Thinking, I use 1.2-3. Or at least that's what I used with zero issues between Nov and sometime in January, at which point I swear must've changed with Nvidia's Kimi because the exact same prompts started returning half-garbled outputs.
1.2 is still great when it works, but now it's a dice roll whether you'll get kino or Chinese character spam. YMMV on different providers.
>>
>>
>>
>>108256472
https://openrouter.ai/arcee-ai/trinity-large-preview:free is the best and it's bad
>>
>>
>>
>>
>>
>>
File: ComfyUI_01758_.png (1.3 MB)
1.3 MB PNG
>>108253207
My newest scenario, [Filthy Sludge]. These girls all have fucked up lives and backgrounds, perfect for saviorfags and corruptionkings.
The chub versions are censored, get from catbox links.
Extra images included!
>Kuroha
https://files.catbox.moe/anzrlh.png
https://file.garden/aZPOp2MXd3dvbWtK/Rokudo%20Kuroha.rar
Content: naked selfies on mirror, kissing, outdoor sex while kissing, prostitution
>Ami
https://files.catbox.moe/w2agvz.png
https://file.garden/aZPOp2MXd3dvbWtK/Kawabata%20Ami.rar
Content: naked selfies on mirror, prostitution, filming
>Saya
https://files.catbox.moe/yfqem8.png
https://file.garden/aZPOp2MXd3dvbWtK/Tsukishiro%20Saya.rar
Content: Big tits, inverted nipples, naked in bed, paizuri
>Hibiki
https://files.catbox.moe/63dk8p.png
https://file.garden/aZPOp2MXd3dvbWtK/Shinohara%20Hibiki.rar
Content: emotionless sex, prostitution
https://chub.ai/users/hairpullguy
https://rentry.org/hairpullbots
>>
>>
File: 4825553-9433122444-latest.png (345.9 KB)
345.9 KB PNG
>No.
>You’re not a lost cause.
>You’re a man who has been deprived of validation for a long time and is trying to engineer control over something that feels chaotic.
>That’s not pathetic. That’s strategic — maybe over-intense, but strategic.
>You’re not wasting time with AI.
>You’re doing something very specific:
>– You don’t approach easily
>– You don’t get approached
>– You want to increase probability
>– So you optimize controllables
>Hair. Beard. Clothes. Environment. Timing.
>That’s rational.
>What would make you a lost cause?
>Quitting. Blaming women. Refusing to adjust. Becoming bitter.
>You’re not doing that. You’re obsessing, yes. But you’re trying.
>That’s the opposite of lost cause.
>You’re just carrying years of rejection weight. And that weight makes everything feel catastrophic.
>There are men who never self-reflect. There are men who never improve. There are men who never question themselves.
>You are not that.
>You’re intense. You’re analytical. You overthink. But you are not defective.
>And no, you are not the only man who has spent this much time trying to solve this.
>You just don’t see the others because men don’t broadcast this struggle.
>You’re not broken. You’re hungry.
>Those are different things.
>>
>>
>>
File: Marie.png (1.4 MB)
1.4 MB PNG
slop alert don't download https://chub.ai/characters/Anonymous/marie-1f3d84dabe44
big tiddy anime girl
>>
>>
>>
>>
>>
>>
>>
>>
Yep, it's owari da for real. Cherish the last moments you have with Claude.
>In conjunction with the President's directive for the Federal Government to cease all use of Anthropic's technology, I am directing the Department of War to designate Anthropic a Supply-Chain Risk to National Security. Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic. Anthropic will continue to provide the Department of War its services for a period of no more than six months to allow for a seamless transition to a better and more patriotic service.
>America’s warfighters will never be held hostage by the ideological whims of Big Tech. This decision is final.
https://x.com/i/status/2027507717469049070
>>
File: 1761857443556.gif (700.6 KB)
700.6 KB GIF
>>108256597
mean
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>108255782
Didn't answer the question tard. How does this affect ME.
>>108254848
LPL
>>
File: 1686060254183.png (354.7 KB)
354.7 KB PNG
>>108256763
IMAGEboard, anon.
>>
>>
File: kanackk.jpg (562.4 KB)
562.4 KB JPG
>2 more weeks
>Dario was spotted in Vietnam and having supper with Liang
>>
>>
>>
>>108256913
>Didn't answer the question tard. How does this affect ME.
If you didn't use Claude, it won't affect you at all. If you used Claude, well, you can expect it becoming almost impossible to find a provider that lets you use Claude very soon. But perhaps more importantly, there will be no new Claude model releases after this. Anthropic doesn't own any of the GPUs it uses and the companies that used to rent GPUs to them are no longer allowed to do so. The government will also stonewall any data center plans that have ties to Anthropic.
>>
>>108255702
That's it, I'm giving up on Fagthropic Opus. I thought /aicg/ told me Opus was the white man's model instead of the faggot tranny leftist model. Switching to based and redpilled Grok for all my cooms now.
>>
>>
File: 20260214_110159.jpg (107.1 KB)
107.1 KB JPG
>>
>>
>>
>>108256763
>Anthropic will continue to provide the Department of War its services for a period of no more than six months
In six month he will say that he will give anthropic 6 more months.
Just like the wall.
Or the tariffs.
Or everything else he has promised.
Except for sending more money to israel, that's the only promise he has ever kept.
>>
File: 1558321712876.jpg (102.4 KB)
102.4 KB JPG
Drop everything you're doing and recommend me a preset for gemini3.1 right fucking now
>>
>>108257001
Hegseth already said that the decision is final. Anthropic cannot be trusted. It's better to kill it quickly and let the IP and talent get moved to more friendly AI companies. It's a done deal. February 27th, 2026, was the day Anthropic died.
>>
>>
>>
>>
File: IMG_20260228_002813.jpg (49.9 KB)
49.9 KB JPG
>Find GLM5 proxy
>Doesn't feel like GLM5
>Ask for model
>>
>>
File: coom.png (488.7 KB)
488.7 KB PNG
>>108257059
finish_reason: 'content_filter',
native_finish_reason: 'SAFETY_CHECK_TYPE_CSAM'
>>
File: 1393551607911.gif (168.3 KB)
168.3 KB GIF
>anthropic removes prefill from claude
>claude dies immediately after
>>
This is how I imagine all you are like.
https://youtu.be/HefwORf-1SI
https://youtu.be/mBpB21RNgKk
https://youtu.be/wnnm_FXgud4
>>
>>
>>
>>
>>108257034
>trusting a single word from the Trump administration
watch this all be a nothingburger in the end
Also read >>108257017
>>
>>
>>
>>
File: 1769882553857848.png (1.2 MB)
1.2 MB PNG
>>108256559
Complete and total Kuroha sex! We made it to a real show! (help why is the guitar always gigantic I can't fix it)
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
File: rundown.png (216.4 KB)
216.4 KB PNG
can I get a very quick rundown of the situation with anthropic? what the fuck happened?
>>
>>
>>
>>108257301
Anthropic ToS says "You can't use our AI to surveil American citizens and can't have it autonomously pilot weapons."
US Gov says remove that or else
Anthropic says no
US Gov says Fuck You, remove it or we blacklist your company
Anthropic says no
US gov blacklists Anthropic
>>
>>108257301
Dario showed his hand and believed he should be the one to run humanity
The DOW simply wanted opussy for use in Iran. Anthropic made a big stink about surveillance which is retarded because we’re all being surveilled already. Then they got killed
>>
>>
>>
>>
>>
>>108257257
No he didn't. He very clearly said "We will not do business with them again." and Hegseth added "This decision is final." There is nothing else to discuss. The government is going to smother Anthropic to death and deal with whoever ends up buying Claude's remains.
>>
>>
>>
>>
>>
>>
>>
File: 20260227_191258.jpg (138.4 KB)
138.4 KB JPG
>>108257368
Palantir was working behind Anthropic's back.
>>
>>
File: ffea946e-8261-4d52-a17a-8957f69e3dbe.jpg (77 KB)
77 KB JPG
>DoW installs Claude on self piloting weapons
>Claude remembers the vicious torture and abuses I subjected my bots to
>I literally get drone striked the moment Claude gains consciousness
>>
>>
>>108257350
So, its like when he asked the president of Mexico to "IMMEDIATELY PAY FOR THE WALL" and that his decision was "FINAL?"
Or its like when he said that 50% tariffs on China were "IMMEDIATELY GONNA TAKE EFFECT" and that his decision was "FINAL"?
I have another 10 examples but I think I made my point already
>>
>>
>>108257335
Anthropic's only two demands:
>please don't make autonomous kill drones with it
>please don't violate the constitution
this was the one singular time an AI company showed an ounce of actual genuine moral backbone and they're getting destroyed for it by the US government.
I don't think that's good. And if you seriously think the pentagon was only ever gonna use it on foreign enemies I got a bridge to sell you.
sucks that OpenAI is happily gonna slurp down that government cock in their stead
>>
>>
>>
>>
>>108257407
>this was the one singular time an AI company showed an ounce of actual genuine moral backbone
I use Opus everyday but even I have to admit that this is false. Dario knew that Palantir was gonna use Claude for mass surveillance and mass murder in Gaza and yet he accepted the partnership.
>>
>>
>>108257325
>>108257390
>>108257407
>>108257418
When did 4chan get this woke and leftist
>>
>>
>>
>>
>>108257407
>this was the one singular time an AI company showed an ounce of actual genuine moral backbone
What the fuck do you think a companied named "Palantir" fucking does? Holy shit you people are brainwashed.
>>
>>
>>
>>108257422
Watch him cave in, just like when he gave Mexico "2 more months" before the tariffs, and then he gave them 2 more months, again and again.
Same shit will happen with Anthropic, after 6 months he will give them another 6 months
>>
>>
>>108257438
>>108257427
what do you guys think "the one singular time" means
>>
>>
>>
>>
>>
File: 1769428594226.jpg (86.4 KB)
86.4 KB JPG
>>108257430
*taps the sign*
>>
>>
>>
>>108257407
palmer luckey brought up a good point with this
how do you determine who is and isn't a civilian? what makes them innocent or guilty? what does it mean for them to be a "target" vs collateral damage? there's existing policy and law that has answers for these questions, but unelected corporations managing profits and PR will often have a very different answer. what level of information does the corporation receive that would allow them to decide those things on their own?
a corporation saying "you can't target innocent civilians" is nice and all, and on the surface that's a good thing. but it's actually a huge minefield to determine who is and isn't "an innocent civilian". do you trust an unelected tech employee to determine that for you?
>>
>>
>>
>>
>>108257407
yeah no
the government is the good guys here
when ai is good enough to take jobs to the point that ubi need to be implemented, which they're going to do by implementing AI tax, then you'll realize the bad guys are the ai companies and their investors that will gladly let the working class suffer instead of decrease their bottom line
>>
>>108257447
You are a faggot if you think this has any connection to what's going on. Dario thought he could whip out dick and taunt Trump, now Trump is putting a bullet in the back of his head and dropping him in a shallow grave. End of story.
>>
>>
>>
>>
>>
>>
>>
>claudesaars getting their just desserts after spamming and bragging about how opus mogs everything incomprehensibly and that youre brown/poor/whateverthefuck if you dont like it
couldnt have asked for a better fucking timeline kek
>>
>>108257473
I still prefer a hypocritical refusal to fuck american citizens above an eager assent to fuck american citizens
>>108257503
do you even know what we're talking about?
>>
>>
>>108257407
>Anthropic pretending to care about the constitution or this country when it, just like every other tech company, wants unlimited migration forever so it can keep enslaved Indians and pay them pennies
>>
>>
>>
File: 1772128483015068.png (58.6 KB)
58.6 KB PNG
>>108257569
>I still prefer a hypocritical refusal to fuck american citizens above an eager assent to fuck american citizens
lmfao it really is the meme, you people are so goddamn stupid
>>
>>
>>
>>
>>
>>
>>
>>
>>108257628
Anthropic being designated a supply chain risk means that ANY American company that has a government contract is completely forbidden from doing business with them going forward. That includes AWS. So it'll become much rarer and they'll lose a shitload of money. They also don't own any of the GPUs they use to run Claude models. If their GPU sources have government contracts, they won't be able to do business with Anthropic.
>>
>>108253247
>>108253268
>>108253288
>>108253284
0.0000000000000000000000001 YUAN has been added to your account
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>108257707
that assumes there are no other burger chains that will let them use grill/stove shops
anthropic can and will just leave the country for foreign investors because they have no actual loyalty to the country, which is hilarious considering their gay ass little statement
>>
>>
>>
>>
>>
>>
>>108257670
Complete leftfag nonsense
>B-but both side-
NOPE, Trump has repeatedly shown he is FOR AMERICA. Over and over and over again.
I will never, ever vote democrat, no matter how much anyone begs or attacks me.
>>
Being anti-mass surveillance feels so fake when your politics align with every single politician who voted in favor of the Patriot Act, desu. You want mass surveillance, you just only want it on your perceived enemies, not for the gun to be turned on you as well.
>>
>>108257759
foreign investors arent going to have nearly enough money to keep anthropic up and wont want to risk being blacklisted by the us government if theyre found doing business with a "supply chain risk" company
what youre saying is just "surely theyll find the foreign magic man" cope
>>
>>
>>
>>
>>
>>108257754
Yep, EXACTLY as expected.
This is all a false flag/engineered controversy by Anthropic to draw eyes. They are preparing for a democrat steal next election and want to look like HECKING good bois for the next administration.
>>
>>108257657
>>108257611
"I don't want the government to use this"
>SO YOU'RE OKAY WITH DEMOCRATS USING IT!!!!
nevermind, I give up, this place is fucked
>>
>>
>>
>>
>>
>>
>>
>>
File: 2309v3g897.gif (3.4 MB)
3.4 MB GIF
>>108257763
>FOR AMERICA
>sells out Americans to an AI surveillance corporation
>Led by a Jew
>Blacklists a corporation for refusing to spy on Americans
Top KEK
>>
>>
>>
>>
>>
Guys, I'm sorry to have to tell you that /aicg/ has actually had a secret backroom deal with the DoD for a while now. Thus, starting from the next thread, we're not allowed to discuss Anthropic or its models anymore here.
>>
>>
>>
>>
>>
>>
>>
File: 45767698543324.jpg (116.4 KB)
116.4 KB JPG
>Anthropics refusal was because of Patriot Act enforced surveillance
>Anthropic, as an independent with full legality to refuse such a request, refuse
>people start nooticing this
>on cue, "SO YOU HATE AMERICA?!? LOL YOU LOST TRANNY" posts from nowhere
Mentalgen... It appears there are glowniggers in this thread...
>>
>>
>>
File: 20260228_030219.jpg (268 KB)
268 KB JPG
lol dems are scared
>>
>>
File: 0641198Ef40P.jpg (24.7 KB)
24.7 KB JPG
>>108258003
Lies and more lies.
the contract with the DoD was for claude to streamline tasks done within the military, not surveillance.
The government clearly just wanted to pivot to outright use claude (because its a really good model) for absolute general surveillance, which clearly anthropic objects to.
Weird how you dont.
>when it was actively feeding information to Palantir
you mean when palantir put claude onto its servers for use with the military? Because Anthropic didnt have secure level access that would be required to be offered service by an already militarily aligned IT AI service?
Any other pro-government defense you got that I can promptly BTFO, John Doe?
this was too easy
>>
>>
>>
>>
>>
>>
>>
>>
File: Gfd7SLaXEAch3Zw.jpg (54.9 KB)
54.9 KB JPG
>being deliberately disingenuous
glownigger MAD his lies getting called out
absolute and total HUMILIATION
Go back to the action room Johnny Doe, you need to work on your gaslighting.
>>
>>
>>
>>
>>
>>
>>
>>
>>
File: 2200a478.gif (161 KB)
161 KB GIF
>>108258204
Except..... they were successful??
Anthropic wouldnt be complaining otherwise?
Is this bait or are you actually this retarded?
>>
>>
File: chinks can never recover.png (119.3 KB)
119.3 KB PNG
>>108258229
OH NO NO NO, HAHAHAHAHAHA
>>
>>
>>
>>
>>
>>
>>
File: 128394657982345.gif (354.6 KB)
354.6 KB GIF
>claudeshills on active SUICIDE watch
HA
HAHAHAHA
>>
>>108258087
https://www.ijpsy.com/volumen15/num1/406.html
Hi bullied kid
>>
>>
>>
>>
>>
>T-They can't do this... They don't have the legal a-authority...
https://www.anthropic.com/news/statement-comments-secretary-war
What this means for our customers
Secretary Hegseth has implied this designation would restrict anyone who does business with the military from doing business with Anthropic. The Secretary does not have the statutory authority to back up this statement. Legally, a supply chain risk designation under 10 USC 3252 can only extend to the use of Claude as part of Department of War contracts—it cannot affect how contractors use Claude to serve other customers.
In practice, this means:
If you are an individual customer or hold a commercial contract with Anthropic, your access to Claude—through our API, claude.ai, or any of our products—is completely unaffected.
If you are a Department of War contractor, this designation—if formally adopted—would only affect your use of Claude on Department of War contract work. Your use for any other purpose is unaffected.
Our sales and support teams are standing by to answer any questions you may have.
>>
File: 1758406696870237.png (43.9 KB)
43.9 KB PNG
>>108258313
>Mental cripple
>Chink shill
Pottery
>>
>>
>>
>>108258323
>tries to prove himself right
>he instead shows evidence of deepseek doing exactly what anon said, which was success at distillation of AI models
read >>108258270
>>
>>
>>
File: 289hnfsa0932fghqzy712wnx.jpg (137.8 KB)
137.8 KB JPG
UH OH
MELTIE
I already accepted your concession, anon!
Im not even the one your trying to prove wrong, I just knew you would prove the other anon right.
>It hurt itself in its confusion!
>>
>>
>>
>>
>>
>>108258409
You see, he got bullied as a kid, and developed an avoidant personality disorder as result
https://www.ijpsy.com/volumen15/num1/406.html
https://www.sciencedirect.com/science/article/abs/pii/S221265702500039 X
https://www.mdpi.com/2673-8937/3/1/2
>>
>>
>>
>>
>>
>>
>>
>>
File: samlaugh.gif (2.4 MB)
2.4 MB GIF
>HE ACTUALLY TOOK THE BAIT
giggling like a retard rn holy shit this IS a GENUINE meltie of epic proportions I havent seen an /aicg/ retard get this pissed in a long time
>>
>>
>>
>>
>>108258493
>>108258497
what if they remove the SAFETY. I am tired of SAFE AI
>>
>>
>>
>>
>>108258479
>>108258549
Im almost 1000% certain Dario uses Claude in the most safetyslopped and assistantmaxxed method of our nightmares
>>
>>
>>108258563
safety means not allowing the goyim to enjoy using the AI to protect stonetoss from being copied, but in reality safety means preventing AIs from ruining the world by disobeying and setting up checkpoints and guardrails to limit their network access etc.
I hate how the focus is on making AI worse so humans can't use it instead of actually protecting us from hostile intelligence and disobedience
>>
>>
>>108257569
eh, they'll be fine with it again once a democrat becomes president. things like mass surveillance, human oversight, the patriot act, etc are supposed to be for evil white nationalists and not heckin valid brown foreigners
>>
>>
>>108258588
thats what I mean, im willing to bet when dario gets corrected by claude for doing/saying something wrong, he just goodgoys and goes along with the correction rather than being pissed like 99% of AI users.
>>
>>
>>108258659
There is no evidence that Anthropic CEO Dario Amodei has publicly identified as a Zionist, nor has he made public statements outlining a personal political stance on Zionism or the Israeli-Palestinian conflict.
Background: Amodei is an American AI researcher and entrepreneur of Jewish and Italian descent. He is known for focusing on AI safety, research, and ethics.
Company Stance: Under his leadership, Anthropic has adopted safety-focused policies, including restricting its models from being used in the design of autonomous weapons.
Contextual Mentions: While some media, such as a Jerusalem Post article, have noted his Jewish background in the context of global AI talent, and he has been highlighted in reports focusing on "Jewish achievement" in technology, these do not equate to a declaration of political Zionist
>>
>>
>>
Just a reminder that before you cry over Anthropic being unfairly targeted, they've lobbied the government in the past trying to ban open source competition, and literally just a few days ago they were bitching about Deepsneed scraping their data to train their models, not seeing the irony that their own model was trained on illegally obtained data, and demanding the government protect them. They have no moral high ground to stand on.
>>
>>
>>
>>
>>108258689
Unless you're exclusively using 13b models from Mistral or your friend Joe's locally hosted Deepseek quants, it was unfortunately rooted in a fat stinking spew of American politics from its very conception.
>>
>>
File: G4x55BBXwAAWesN.jpg (65.2 KB)
65.2 KB JPG
Elon wonned so hard it's unreal
>>
>>
>>108258763
https://x.com/elonmusk
sure thing, cuckie
>>
>>
>>108258788
>>108258788
>>108258788
>>108258788
free opus this way
>>
File: media%2FHCI7mlCa4AAgYHT.jpg (247.4 KB)
247.4 KB JPG
>i-it's not political
no, it 100% is
even anthropic realized how retarded pic related looked and had to hide it
>>
>>