Jul 18
“비싼 Max 요금제를 쓰는데 최고급 모델은 따로 결제해야 하나?” Claude Fable 5가 2026년 7월 20일부터 Max와 Team Premium 요금제에 사용 한도의 50%까지 기본 포함된다. Pro·Team Standard의 100달러 크레딧과 출시 중단부터 복구까지의 과정을 정리했다. #AI코딩 #Anthropic #Claude #ClaudeMax #Fable5 #생성형AI Read the full article
seen from Germany
seen from United States
seen from Netherlands
seen from Sweden
seen from United States
seen from Sweden
seen from Singapore
seen from Poland
seen from United States
seen from Nepal
seen from United States
seen from United States
seen from United Arab Emirates
seen from Singapore
seen from Singapore
seen from Netherlands
seen from Germany
seen from Sweden
seen from United States
seen from United Kingdom
“비싼 Max 요금제를 쓰는데 최고급 모델은 따로 결제해야 하나?” Claude Fable 5가 2026년 7월 20일부터 Max와 Team Premium 요금제에 사용 한도의 50%까지 기본 포함된다. Pro·Team Standard의 100달러 크레딧과 출시 중단부터 복구까지의 과정을 정리했다. #AI코딩 #Anthropic #Claude #ClaudeMax #Fable5 #생성형AI Read the full article
Sony Kills the Disc, Anthropic's Fable 5 Is Back, and South Korea Bets $576B That Chips Are Everything
Rough week to be a physical-media hoarder. Sony dropped a bombshell on July 1 — it's pulling the plug on PlayStation game discs starting January 2028. That's not that far off. Meanwhile Anthropic's Fable 5 and Mythos 5 just got unbanned after an 18-day US export control standoff, and South Korea unveiled a jaw-dropping $576 billion semiconductor strategy. Plus GM locked down a direct chip deal with Micron, and there's a new AI inference breakthrough that might actually make your chatbots faster without wrecking your wallet. Let's dig in.
🎮 Sony Finally Draws the Line — Discs Are Done by 2028
Sony made it official on July 1: starting January 2028, no more physical game discs for new PlayStation titles. That means PS6 is almost certainly going digital-only, and PS5's disc-drive version might be the last of its kind. Honestly, this was coming. PC gaming already proved digital distribution works. But for the physical collectors, the folks who actually trade in games or buy used — this stings. Rockstar already skipped a disc for GTA VI, so the writing was on the wall.
From a practical standpoint: if you're sitting on a PS5 with a disc drive, you've got about 18 months of new disc releases left. After that? eBay prices for physical copies are gonna climb. One thing that bugs me though — Sony didn't say anything about backward compatibility for existing discs. If PS6 can't play my PS4/PS5 discs, that's a real middle finger to anyone with a shelf full of games.
🤖 Anthropic's Fable 5 Is Back Online — With Strings Attached
On June 30, the Trump administration quietly lifted export controls on Anthropic's Claude Fable 5 and Mythos 5 models. These were banned on June 12 over national security fears — the idea that state actors could weaponize frontier AI. The ban effectively cut off global users. Now Anthropic's slowly restoring access with a 50% weekly usage cap through July 7 and new safety classifiers in place.
To be fair, the ban itself was kind of chaotic. The US Commerce Department dropped it without much public explanation, then reversed course after 18 days. That's not exactly reassuring if you're a developer building on top of these models. Anthropic's now operating under formal government oversight, which sounds like a polite way of saying "we're being watched." For regular users? Fable 5 is genuinely impressive — I've been testing it on coding tasks and it's noticeably sharper than Claude 4 on long-context prompts. Whether that's worth the regulatory drama is another question.
🧠 South Korea's $576B Chip Plan — Samsung and SK Hynix Take Center Stage
South Korea dropped what might be the biggest single-country semiconductor investment package in history: $576 billion focused on AI chips, memory manufacturing, and regional industrial hubs. Samsung and SK Hynix are the main drivers. This isn't just about keeping up with TSMC — it's about owning the entire AI memory pipeline. HBM (high-bandwidth memory) is the bottleneck for every AI data center right now, and SK Hynix and Samsung make basically all of it.
Quick add-on note: this kind of spending is why DRAM and NAND prices are climbing. If you're planning a PC build this year, buy your RAM and SSD sooner rather than later. The Longsys earnings report confirms it — China's biggest memory module maker just reported a 600-fold profit jump. That's not a typo. Six hundred times. Tight wafer supply plus AI-driven demand equals expensive memory across the board.
🚗 GM Goes Direct to Micron for Automotive Chips
GM signed a long-term agreement with Micron to secure memory and storage chips directly — bypassing the spot market that's been running hot since the pandemic. This is a smart move for two reasons. One: cars are basically computers on wheels now, and every infotainment system, ADAS camera, and EV battery controller needs chips. Two: GM got burned during the 2021-2022 shortage and clearly doesn't want a repeat. The partnership also includes joint innovation on future automotive tech, which could mean custom memory solutions down the road.
For consumers? This probably won't make cars cheaper — but it might prevent the nightmare scenario where you can't get a new Chevy because some random sensor chip is backordered 14 months.
⚡ STAR-KV: Making AI Inference Suck Less
Dnotitia, a Korean AI infrastructure startup, just published STAR-KV — a low-rank KV cache compression technique that was selected as a spotlight paper at ICML 2026 (that's about 2.2% of submissions, for context). The headline number: up to 20x compression on the KV cache with only minimal quality loss. In practice, that means attention computation speeds up by 6.9x and overall generation throughput improves by 3.1x. What does that mean for someone using ChatGPT or Claude? Faster responses, longer context windows without hitting a memory wall, and potentially cheaper API costs since providers need fewer GPUs per query.
The paper and code are on arXiv and GitHub if you're into the technical weeds. But the takeaway for the rest of us: AI inference is getting real efficiency improvements, not just bigger models. That's the kind of progress that actually matters for day-to-day users.
🎬 Quick Hits
Netflix is using an AI-generated Gene Wilder voice (via ElevenLabs, with family consent) for its upcoming Wonka's The Golden Ticket reality show. Creepy? A little. Legally cleared? Apparently yes. Bending Spoons — the Italian company that bought AOL, Evernote, and Vimeo — is going public at a potential $19 billion valuation. And TV Time, the popular tracking app, is shutting down July 15 as its parent company pivots to enterprise AI. Feels like every consumer app with half a user base is chasing AI dollars these days.
Alright, where does that leave us?
If you're building a PC right now, pull the trigger on storage and RAM — prices are only going one direction through the end of the year. For gamers, the PS5 disc version is still a safe buy if you want to play physical through 2028, but don't expect the next console to have a drive. And for anyone using AI tools day-to-day, keep an eye on inference efficiency improvements — they'll quietly make your experience better without you noticing, which is honestly the best kind of tech upgrade.
What's your take on Sony going all-digital? Is losing physical discs a real loss or just nostalgia talking? Drop your thoughts below.
More reads → PayCalc at k232323.sbs
Sony Kills Physical Discs, Anthropic's Fable 5 Returns, and Super Micro Gets Hit with Export Heat That's this week in tech — and honestly, it's been one of those weeks where the news cycle just doesn't let up. We've got regulatory drama around AI models, a new memory trick that could actually make LLMs cheaper to run, some serious legal trouble for Super Micro's Taiwan arm, and Sony dropping a bomb on physical game collectors. Let's dig in. --- **Anthropic's Fable 5 is back — but the export control genie isn't going back in the bottle** The biggest AI policy story this week is Anthropic getting the green light to restore Fable 5 after the Trump administration lifted those export restrictions. For context, Fable and Mythos were hit with a sudden suspension back in June over national security concerns — specifically, the worry that they could be weaponized by bad actors. Now they're back with new safeguards and monitoring layers baked in. To be fair, this whole episode tells you something about where AI is headed. Model launches aren't just product announcements anymore — they're geopolitical events. The Verge reported that Anthropic had to go through serious negotiations with the government, adding safety classifiers, vulnerability reporting pipelines, and government access protocols. My take? This is going to become the norm for frontier models. If you're building on top of Claude or GPT, expect more of these stop-and-go cycles depending on who's in office. Quick add-on note — OpenAI is reportedly floating the idea of giving the U.S. government a 5% stake in the company. First reported by Financial Times, the proposal is apparently being discussed between Sam Altman and administration officials. *Unverified industry rumor, for reference only.* A weird idea on paper, but it signals something real: frontier AI companies are starting to look more like public utilities than private software firms. Whether that's good or bad probably depends on how much you trust the government to have a seat at the strategy table. **Dnotitia's STAR-KV: 20x compression on LLM memory — and this one's actually peer-reviewed** Here's one I haven't seen getting enough attention outside the ML crowd. Dnotitia, a Korean AI infrastructure startup, released STAR-KV — a low-rank approach to KV cache compression that just got selected as a Spotlight paper at ICML 2026. That's the top 2.2% of submissions, so it's not vaporware. The headline number: up to 20x compression on KV cache, which is one of the main bottlenecks when you're running long-context inference. They're claiming attention computation speeds up by 6.9x and overall generation throughput by 3.1x. The source code is already on GitHub, which is refreshing — no paper-only mystery sauce. Honestly, this is the kind of work that moves the needle for anyone running self-hosted models. If you're running Mistral or Llama locally and hitting memory walls on long documents, techniques like this are what'll make 128k context actually usable on consumer hardware. Keep an eye on this one. **Super Micro's Taiwan unit in legal hot water over AI server exports** Now onto the hardware side. Super Micro's Taiwan division had four employees detained on allegations of illegally exporting AI servers packed with Nvidia chips — allegedly bypassing US export controls. Taiwanese authorities are investigating document forgery and breach of trust, and this is getting pretty ugly. This one hits close to home for anyone building AI infrastructure outside the US. The Super Micro case shows that export control enforcement is tightening — and not just at the chip level. Entire server systems are getting scrutinized. If you're running a smaller AI lab or colo operator outside the US, this might mean longer lead times and pricier hardware procurement. **Micron and GM lock in a chip deal — your next car runs on DRAM** Micron and General Motors signed a long-term semiconductor supply agreement this week. It's not just about memory for infotainment screens — modern vehicles are soaking up DRAM and NAND for ADAS, autonomous driving compute, and over-the-air update systems. GM wants to lock in supply after the 2021-2023 chip shortage nightmare. Parallel story: Chinese automakers are pushing for 100% domestic chips in their vehicles by 2027. Crypto Briefing reported that Beijing's carmakers are trying to eliminate foreign semiconductors entirely. That's a $100B+ market shift in the making. For North American and European buyers, this could mean more regional supply chain fragmentation — and potentially higher prices on imported cars with sensitive chip content. **Sony is killing physical PlayStation discs — 2028 is the deadline** This one hits the gut if you're a collector. Sony confirmed it will stop producing physical discs for all new PlayStation games starting in 2028. TechCrunch broke the story, and it's the clearest signal yet that the all-digital future is here — whether we asked for it or not. Personally, I get it from a logistics standpoint. Physical media is expensive to manufacture, ship, and retail. But losing the ability to buy, sell, or trade used games is a bummer. For PC gamers this is old news — we've been all-digital for years. But console gamers might want to hold onto those disc editions while they still work. **Quick hits I don't want to skip** - **Apple's Hide My Email has a bug** that a researcher claims exposes your real email address — rendering the feature useless. Apple hasn't confirmed the fix timeline. If you rely on this for signups, maybe go manual for a bit. (TechCrunch) - **Bending Spoons went public** — the Italian company behind AOL, Evernote, and Vimeo IPOs at a $19B valuation. NYT covered it. It's one of the weirder tech IPOs this year given their portfolio is basically resurrected internet corpses, but the market seems to love the turnaround narrative. - **National Grid dropped $1.75B on a 35% stake in Joulent**, a US energy platform building power for AI data centers. First project: a 2.67GW gas plant in West Texas tied to Microsoft. AI's power hunger is real, and utilities are cashing in. - **Netflix is using an AI-generated Gene Wilder voice** for its Willy Wonka reality show, with family consent via ElevenLabs. The ethics of this are... a conversation for another day. But technically impressive. --- Look, this week's lineup confirms something I've been saying for a while — tech in 2026 is less about shiny launches and more about who controls the pipes. Model access, chip supply chains, power grids, export rules, even physical game discs. The gold rush phase is giving way to the consolidation phase. If you're building a rig this quarter: hold off on the RTX 5090 Super unless you absolutely need it — prices are still inflated from AI demand pulling the same wafers. For office builds, DDR5 is finally getting reasonable, and last-gen Ryzens are a steal right now. Console gamers? I wouldn't panic-buy physical copies just yet, but maybe grab a disc drive PS5 if you care about resale. What's catching your attention this week — the Fable 5 unblock, the Sony disc shutdown, or something I missed? Drop a comment, always curious what the community is watching. P.S. If you're still planning your week and need a solid calculator tool, I've been using Decision Calculator — handles everything from build budgets to PSU wattage estimates. Nothing fancy, just gets the job done.
Claude Fable 5 returns after a three-week suspension. See what's new, coding changes, and if Claude Code users should switch.
Claude Fable 5 returns after a three-week suspension. See what's new, coding changes, and if Claude Code users should switch.
Anthropic ile Trump Yönetimi Arasındaki Gerilim Kime Yarıyor?
Yapay zekâ şirketi Anthropic ile Trump yönetimi arasındaki gerilim son haftalarda teknoloji sektörünün en çok konuşulan konularından biri haline geldi. Şirketin en yeni modelleri olan Fable 5 ve Mythos 5'in ihracat kontrolü gerekçesiyle erişime kapatılması, yalnızca yapay zekâ güvenliği tartışmalarını değil, siyasi etkileri ve rekabet dengelerini de yeniden gündeme taşıdı.
Tartışmalı Kararın Ardındaki Sorular
ABD yönetimi tarafından gönderilen resmi talepte, Anthropic’in gelişmiş modellerinin yabancı uyruklu kişiler tarafından kullanılmasının engellenmesi istendi. Ancak karara gerekçe gösterilen ulusal güvenlik riskleri kamuoyuyla paylaşılmadı.
Bu durum teknoloji çevrelerinde çeşitli soru işaretleri doğurdu. Bazı uzmanlar kararın teknik güvenlik kaygılarından çok, Anthropic ile Trump yönetimi arasında uzun süredir devam eden anlaşmazlıkların bir sonucu olabileceğini öne sürüyor.
Özellikle şirketin geçmişte devletin bazı taleplerine karşı mesafeli durması ve çeşitli politika konularında Washington ile ters düşmesi, sürecin yalnızca teknik bir mesele olmadığını düşündüren unsurlar arasında gösteriliyor.
Rakipler İçin Avantaj mı?
Kararın ardından gözler doğal olarak Anthropic’in rakiplerine çevrildi. Yapay zekâ pazarında faaliyet gösteren diğer büyük şirketlerin benzer yaptırımlarla karşılaşıp karşılaşmayacağı merak konusu oldu.
Bazı analistler, mevcut durumun rakip şirketler için kısa vadeli bir avantaj yaratabileceğini düşünüyor. Ancak diğer taraftan, düzenleyici kararların şirketler arası ilişkilere bağlı olarak şekillenmesi ihtimali sektör açısından daha büyük bir belirsizlik yaratıyor.
Teknoloji şirketleri için asıl risk, hangi kuralların uygulanacağından çok bu kuralların ne kadar öngörülebilir olduğu olarak görülüyor.
Siber Güvenlik Uzmanlarından Tepki
Karar yalnızca yapay zekâ sektöründen değil, siber güvenlik topluluğundan da eleştiri aldı.
Çok sayıda güvenlik uzmanı tarafından imzalanan açık mektupta, gelişmiş güvenlik yeteneklerine sahip modellerin erişime kapatılmasının ABD’deki savunma ve ağ güvenliği ekiplerini olumsuz etkileyebileceği savunuldu.
Uzmanlara göre bu tür modeller yalnızca potansiyel tehditler oluşturmakla kalmıyor, aynı zamanda güvenlik açıklarının tespit edilmesi ve savunma sistemlerinin güçlendirilmesi için de önemli araçlar sunuyor.
Anthropic İçin Beklenmedik Bir Kazanç Olabilir
İlginç şekilde yaşanan kriz, Anthropic’in marka algısına zarar vermek yerine şirketin görünürlüğünü artırabilir.
Geçmişte şirket ile Trump yönetimi arasında yaşanan başka anlaşmazlıkların ardından Anthropic ürünlerine olan ilginin arttığı biliniyor. Özellikle Claude platformunun kullanıcı sayısında önemli yükselişler görülmüş, şirket yapay zekâ dünyasında daha fazla dikkat çekmeye başlamıştı.
Sektör uzmanları, "fazla güçlü olduğu için kısıtlanan model" algısının bazı kullanıcılar ve geliştiriciler açısından cazip bir pazarlama etkisi yaratabileceğini belirtiyor.
Güçlü Model Söylemi İki Ucu Keskin Kılıç
Anthropic uzun süredir gelişmiş yapay zekâ sistemlerinin oluşturabileceği risklere dikkat çeken şirketlerden biri olarak öne çıkıyor. Ancak şirketin bir yandan yapay zekânın tehlikelerine vurgu yaparken diğer yandan son derece güçlü modeller piyasaya sürmesi de eleştirilerin odağında yer alıyor.
Bazı yorumculara göre şirket, güvenlik uyarıları ile teknolojik ilerleme arasında hassas bir denge kurmaya çalışıyor. Ancak bu yaklaşım, modeller üzerindeki kamuoyu ve devlet denetimini de artırıyor.
Yapay Zekâ Savaşlarında Yeni Bir Dönem
Anthropic ile Trump yönetimi arasında yaşanan son gerilim, yapay zekâ sektöründeki rekabetin artık yalnızca teknoloji şirketleri arasında yaşanmadığını gösteriyor.
Ulusal güvenlik, ihracat kontrolleri, düzenleyici baskılar ve siyasi ilişkiler, yapay zekâ şirketlerinin geleceğini şekillendiren faktörler arasında giderek daha fazla önem kazanıyor.
Kısa vadede Anthropic bazı ürünlerini geri çekmek zorunda kalmış olsa da, yaşananların şirketin pazar konumunu zayıflatmak yerine daha fazla görünürlük ve ilgi getirmesi ihtimali sektör tarafından yakından izleniyor.
El G7 teme que EE.UU. apague la IA americana de un día para otro
Macron y Modi expresaron en la cumbre del G7 su preocupación porque Washington pueda bloquear el acceso a modelos de IA estadounidenses sin previo aviso. El bloqueo a Fable 5 y Mythos 5 de Anthropic dejó de ser un temor hipotético y se convirtió en advertencia concreta para gobiernos y empresas de todo el mundo (Fuente Techcrunch). El apagón de los modelos más avanzados de Anthropic dejó de ser…
¿Por qué EEUU bloqueó Fable 5 tras el prompt «arregla este código»? #Fable5 #Anthropic #InteligenciaArtificial #IA #Ciberseguridad #Jailbreak #Tecnologia #EstadosUnidos #Trump #SeguridadNacional #Codigo #Innovacion #felizmiercoles #17dejunio
Fable 5 e la Sfida Globale sulla Sicurezza dei Modelli Artificiali
Fable?5 e la crisi globale sulla sicurezza dei modelli di IA
*Settimana turbolenta* - Antropic ha lanciato due nuovi modelli di frontiera, Fable?5 e Mythos?5, annunciati come un "quarto livello" rispetto a Iku, Sonnet e Opus. Sono progettati per compiti complessi a lungo termine, operando in modo autonomo e offrendo memoria contestuale potenziata, capacità visive avanzate e notevoli miglioramenti in programmazione, knowledge work e analisi documentale.
### 1. Prestazioni e costi | Aspetto | Fable?5 vs concorrenti | |---------|------------------------| | Programmazione (Sweet?Bench/Frontier?Code) | 80.3?pt, +?30?pt rispetto a Gemini (Google) | | Knowledge work | 13.3?pt vs 2.1?pt di GPT?5.5 | | Memoria | 3× rispetto a Opus?4.8 | | Visione (PDF lunghi, tabelle, grafici) | Allucinazioni drasticamente ridotte | | Costi API | ~2× Opus?4.8, consumo energetico elevato | | Infrastruttura | Supportata da hardware "Colossus" (Elon?Musk) |
> La potenza di calcolo è tale da richiedere investimenti ingenti; l'uso "parsimonioso" è consigliato per contenere le spese.
### 2. Sistemi di sicurezza "Safeguard" - Classificatori automatici filtrano in tempo reale richieste per settori ad alto rischio (chimica, biologia, sicurezza informatica). - Falsi positivi: in ~5?% delle sessioni una richiesta ambigua viene declassata; il lavoro passa a Opus?4.8 ma il cliente paga comunque il prezzo di Fable?5, senza trasparenza sul downgrade. - Protezione contro la "distillazione": tentativi di estrarre le capacità del modello vengono bloccati e reindirizzati a versioni meno avanzate. - Obiettivo: bilanciare la ricerca legittima con il rischio di exploit, mantenendo la proprietà intellettuale di Antropic.
### 3. Intervento governativo USA - Blocco immediato a tutti gli utenti non cittadini statunitensi (inclusi dipendenti stranieri di Antropic) a causa di una presunta vulnerabilità. - Progetto Glasswing: fase preliminare riservata a partner di difesa cibernetica e a esperti del AI Safety Institute (UK). L'estensione globale è stata fermata dal governo. - Nuova policy di data?retention: tutti i log delle interazioni sono conservati per 30?giorni per audit e sicurezza nazionale; i dati non saranno ri?addestrati ma saranno analizzati dalle autorità.
### 4. Retroscena del blocco (90?minuti) - L'ordine di Washington ha concesso a Antropic solo 90?minuti per chiudere l'accesso globale, tempo ritenuto insufficiente per una piattaforma distribuita a milioni di utenti. - La notizia è stata amplificata da una segnalazione interna di Amazon (principale finanziatore di Antropic) che ha evidenziato un jailbreak capace di bypassare le protezioni del modello su larga scala. - Antropic ha contestato il provvedimento, sottolineando che le capacità di individuare bug sono essenziali e già disponibili in altri modelli (es. GPT?5.5). - Il caso ha generato dibattiti su: * Rischi reali vs hype di sicurezza cibernetica. * Possibile strategia di marketing in vista dell'IPO di Antropic. * Precedenti tensioni con il Dipartimento della Difesa su usi militari dell'IA. * Impatto di future "sospensioni a sorpresa" sull'innovazione IA a livello globale.
### 5. Conclusioni - Fable?5 rappresenta un salto significativo in potenza e versatilità, ma con costi operativi elevati e un'impronta energetica notevole. - I Safeguard introducono una barriera di sicurezza senza precedenti, ma generano frustrazione per i falsi positivi e per la mancanza di chiarezza sui downgrade. - L'intervento degli USA evidenzia il crescente ruolo delle autorità nazionali nella regolamentazione dei modelli di IA strategica, creando una nuova "geopolitica dell'IA" dove la cittadinanza diventa criterio di accesso. - La rapidità del blocco ha mostrato la vulnerabilità delle catene di distribuzione di IA avanzata e ha sollevato preoccupazioni su possibili futuri garanzie punitive da parte dei governi, con potenziali effetti di "freezing" sull'avanzamento tecnologico.
In sintesi, la vicenda di Fable?5 riflette il delicato equilibrio tra innovazione potenzialmente rivoluzionaria, costi economici e energetici, sicurezza nazionale e politiche di controllo geopolitico che sta ridefinendo il panorama dell'intelligenza artificiale a livello mondiale.