Vibekollen Nytt
Hämtar senaste nytt…
Vibekollen Nytt
Modeller, verktyg och händelser som formar AI-världen. Hitta nyheter, video och poddar – och följ vad som händer på Vibekollen.
Alla nyheter i tidsordning. Använd filtren för att hitta ditt ämne, din podd eller ditt verktyg.
TelegramTechmeme
Today's Stratechery Interview is with Palantir CTO Shyam Sankar. Sankar has worked at Palantir for 20 years …
29 okt. stratechery.com
TelegramTechmeme
Circuit Mind, a startup developing AI software for electronics design, raised $25 million, CEO Tomide Adesanmi tells Axios Pro exclusively.
29 okt. axios.com
BloggMistral AI
The weights of Mistral Large 4 (Le Chonk) are now open. A 1T parameter multimodal model, and best open-weight model for coding, trained in Europe.
29 okt. mistral.ai
BloggCursor
Today we identified a security incident involving a specific extension package in the Cursor VS Code extension Marketplace: Anthropic.claude-code version 2.1.313.
29 okt. cursor.com
BloggGoogle Gemini
Turn raw notes into high-impact, ATS-friendly resume bullets using Gemini. Follow these four steps to polish your professional profile. Get the details.
29 okt. blog.google
TelegramTechmeme
OpenAI's GPT-6 Astra, the most powerful model the company has released, has met with strong demand since its release in early September.
29 okt. theinformation.com
TelegramTechmeme
In Brief: Following CEO Tim Sweeney's announcement of a “huge pivot” towards AI tools, Epic Games has confirmed that AI tools …
29 okt. thisweekinvideogames.com
TelegramTechmeme
Sheryl Sandberg says the 32-year-old executive reminds her of Mark Zuckerberg. Josh Payne must now take his startup public at a tumultuous moment for the AI industry.
29 okt. bloomberg.com
VideoAI Engineer
David Abitbol, who leads AI operations at Revolut, shares data from 2,000 engineers on what AI tools did to coding time and lead time, and why self-hosting open-source GLM models made the biggest difference: about 60-70% faster delivery at roughly a tenth of what they were paying frontier labs. He covers the GPUs needed, the latency targets that keep engineers happy, the tuning that made the same hardware 88% faster, and how an EU-regulated bank keeps it compliant. Chapters: 0:00 Intro 0:36 Who Revolut is 1:18 What AI did to 2,000 engineers' lead time 2:45 Why they started self-hosting 3:33 Coding isn't the bottleneck 4:24 What it takes to self-host GLM 5:46 Tuning: same GPUs, 88% faster 7:06 The parameters that matter 8:14 Production setup 9:23 For the CIO: compliance and safety David Abitbol https://www.revolut.com Join us in SF for the second annual AIE CODE summit: https://aie.codes our last flagship summit of 2026. Applications live now!
29 okt. youtube.com
BloggOpenAI
OpenAI uses layered monitoring to detect frontier model misalignment and unauthorized agent behavior so responders can intervene before incidents escalate.
29 okt. openai.com
PoddPractical AI
In this special episode, we bring you highlights from three talks at the Midwest AI Summit. Heidi Porter explores how to build reliable AI-augmented workflows using the right tools, evaluations, and guardrails. Kevin Solorio shares why building AI with frontline workers, not just for them, is key to earning trust and driving adoption, emphasizing collaboration and continuous feedback. And Chris Shinkle and Andrew Dunlap of SEP discuss how software teams can find the right level of AI autonomy based on their goals, constraints, and verification needs, sharing lessons from teams using AI to build software more effectively.Featuring: Heidi Porter – LinkedInKevin Solorio – LinkedInChris Shinkle – LinkedInAndrew Dunlap – LinkedInDaniel Whitenack – Website, GitHub, XLinks:Moyer Group AISEPFocusedSponsors:Hostinger: Create your website today.
29 okt. share.transistor.fm
VideoAI Engineer
Temporal's State of Development survey found that 80.8% of engineers use agents daily, but only 49.1% run them in production. Developer advocate Shy Ruparel walks through the four gaps behind that number (usage vs. production, trust vs. reality, prototype vs. enterprise, seeing vs. recovering), then demos a $2,500 payout workflow with a local model that survives having its workers killed mid-flight, without losing state, repeating tool calls or re-spending tokens. Chapters: 0:00 Intro 0:34 Finance runs on long-running workflows 2:05 The State of Development survey 3:16 Gap 1: Usage vs. production 4:33 Gap 2: Trust vs. reality 7:18 Gap 3: Prototype vs. enterprise 8:55 Gap 4: Seeing vs. recovering 10:53 Demo: a payout that survives a crash 13:06 Durable execution explained 18:12 What the demo solved 20:10 Who runs this in production 20:50 Wrap-up Shy Ruparel https://temporal.io Join us in SF for the second annual AIE CODE summit: https://aie.codes our last flagship summit of 2026. Applications live now!
29 okt. youtube.com
VideoAI Engineer
Code generation is moving faster than ever, and the bottleneck has moved to review. Cole Gannaway, a solutions engineer at Qodo, shows how a one-line data model change that passes every test can quietly break web, mobile, CLI and MCP clients, and how a software graph of your repos reveals that blast radius. He covers turning tribal knowledge and PR comments into custom rules, bringing review context in before code is written, why the agent that wrote the code shouldn't grade it, and checking PRs against the original ticket. Chapters: 0:00 Intro 0:59 Review is the new bottleneck 2:03 A one-line change that passes every check 3:35 The hidden blast radius 4:15 Building a software graph 4:56 Custom rules and tribal knowledge 5:47 Context before code is written 6:29 Don't let the writer grade its own code 8:20 Turning PR comments into rules 10:06 Q&A: catching the wrong intent 13:55 Wrap-up Cole Gannaway https://www.qodo.ai Join us in SF for the second annual AIE CODE summit: https://aie.codes our last flagship summit of 2026. Applications live now!
29 okt. youtube.com
BloggOpenAI
Ropes & Gray is developing AI for transactional diligence with OpenAI. Lawyers review grouped documents, summaries, and reports before advising clients.
29 okt. openai.com
TelegramTechmeme
When Pony Ma called Tencent Holdings Ltd.'s AI endeavor “a leaky ship” this summer, the billionaire founder sent ripples through …
29 okt. bloomberg.com
TelegramTechmeme
Computer and information science departments at US colleges saw an exodus of students this fall, with enrollment declining more than 12% …
29 okt. bloomberg.com
TelegramTechmeme
That's a problem, cos it just spent $51 billion on infrastructure and still needs to rent compute from others to run its cloud
29 okt. theregister.com
TelegramTechmeme
Procurement startup Fragment is exiting stealth with $56 million raised, CEO Pranav Mulgund tells Axios Pro exclusively.
29 okt. axios.com
TelegramTechmeme
The iPhone maker for years bullied flash-storage makers into steep discounts—until AI hyperscalers bid up prices and gobbled up supply
29 okt. wsj.com
