Good morning. It’s Wednesday, August twentieth.
On at the present time in tech history: In 1964Berkeley’s Project Genie quietly modified the operating system game. Funded by ARPA, it introduced memory paging, protected user modes, and interactive time-sharing, concepts that seeped into TENEX and later Unix. Ken Thompson, before co-creating Unix at Bell Labs, actually frolicked hacking on one among these machines.
You read. We listen. Tell us what you think that by replying to this email.
Turn your rough idea into a completely functional agent!
Nuvi will guide you each step of the way in which.
-
No code, no latest tool to learn—just a straightforward document you edit with the assistance of Nuvi assistant
-
Quick iteration—change the doc → try again → improve in minutes
-
One-click setup—connect your tools, share together with your team
-
AI BDR Agent (Gmail, Google Sheets, HubSpot)
-
RFP generator (SharePoint, Perplexity)
-
Marketing Assistant (Reddit, X, Google Sheets)
Thanks for supporting our sponsors!
Today’s trending AI news stories
DeepSeek V3.1 just dropped — and it could be essentially the most powerful open AI yet
DeepSeek has quietly dropped V3.1, a 685B-parameter open-source model now live to tell the tale Hugging Face. The update expands the context window to 128k tokens (roughly a 300-page book) and introduces a hybrid architecture that merges chat, reasoning, and coding in a single system. Researchers digging through the discharge found hidden tokens enabling search integration and internal reasoning, hinting at latest architectural experiments.
DeepSeek V3.1 hits token efficiency once reserved for closed models — cutting cost and latency head-on.
On benchmarks, V3.1 posted a 71.6% rating on the Aider coding test, topping all Chinese systems and beating Anthropic’s Claude Opus 4 in cost-efficiency, delivering ~$70 workloads for about $1. The model supports BF16 and FP8 precision, making it tunable for various hardware setups, though its 700GB size keeps it compute-heavy.
The update also got here with the quiet removal of references to DeepSeek’s R1 reasoning model, raising questions on delays to the next-gen R2, reportedly stalled by Huawei Ascend chip issues. Read more.
Sam Altman on GPT-6: ‘People want memory’
OpenAI is already looking beyond GPT-5, with CEO Sam Altman teasing GPT-6 as a faster, more adaptable model designed to recollect user preferences, routines, and tone, and permit fully personalized assistants. Memory is the feature Altman calls his favorite of 2025, though temporary memory isn’t yet encrypted, raising privacy concerns for sensitive legal or medical queries. GPT-6 can be being designed for compliance with a U.S. executive order mandating government AI remain ideologically neutral but customizable, and Altman envisions integration with neural interfaces, robotics, and novel compute substrates.
Meanwhile, In Microsoft and Edinburgh’s OdysseyBench, a 602-task, multi-day test across Word, Excel, PDFs, email, and calendars, OpenAI’s older o3 model topped the newer GPT-5 on the core thing that matters: sustained, cross-app execution. On the toughest OdysseyBench-Neo split, o3 hit 61.26% vs. GPT-5 at 55.96% When workflows required three apps without delay, o3 scored 59.06% to GPT-5’s 53.80%.
OdysseyBench includes each easy, single-step tasks and complicated, long-term office workflows that require models to administer conversations and coordinate multiple applications. | Image: Wang et al.
The catch is that each models still fumble on long-horizon planning. It botched DOCX/XLSX edits, skipped steps, used flawed tools – precisely the brittle edges that break real office automations. Today’s agents may reason higher, but reliability across multi-step, multi-app sequences stays the bottleneck.
On the adoption front, OpenAI launched ChatGPT Go in India at $5/month (₹399), doubling memory, expanding quotas for images, file uploads, and Python evaluation, and offering tools for projects, tasks, and custom GPTs. Read more.
Meta drops DINOv3 and opens Vision AI for industrial deployment
DINOv3, Meta’s latest self-supervised vision model, has been open-sourced . Seven billion parameters, 1.7 billion images, and no labeled data needed, Meta is giving developers all the things: pre-trained variants, adapters, and full training pipelines. The result is powerful generalization across messy, annotation-poor domains like satellite imagery or medical scans, and the power to drop it straight into production as an alternative of just experimenting within the lab.
Meta has also reportedly overhauled its AI org, consolidating teams into 4 groups under Meta Superintelligence Labs. TBD Labs, run by Alexandr Wang, will give attention to foundation models like Llama, while the opposite units handle research, product integration, and infrastructure. It’s a decent setup geared toward accelerating development and staying consistent with OpenAI, Anthropic, and Google DeepMind.
On the user side, Meta AI now auto-translates Instagram and Facebook reels between English and Spanish, syncing dubbed audio to lip movements. Using neural machine translation and generative audio modeling, the system preserves voice expression and routinely surfaces localized content to users, bridging language barriers while maintaining engagement. Read more.
ElevenLabs launches Chat Mode to let businesses construct text-only AI agents
ElevenLabs has upgraded its Conversational Agent platform with Chat Mode, letting AI flip effortlessly between text and voice depending on user context. The system analyzes device type, environmental noise, and user behavior, hitting 85% accuracy in real-time mode selection during beta tests.
Introducing Chat Mode
You’ll be able to now construct text-only conversational agents.
Ideal for:
– Customers that prefer typing to speaking.
– Precise inputs like order IDs or email addresses.
– Solving easy issues, handing off to our voice agents for complex tasks.— ElevenLabs (@elevenlabsio)
3:00 PM • Aug 19, 2025
Leveraging cloud infrastructure and edge computing, response latency drops below 200ms, and low-code APIs enable deployment inside days across CRM and workflow systems. Chat Mode enhances customer interactions, from support to sales, while integrating safeguards against bias and deepfake misuse. By enabling context-aware, multi-modal communication, ElevenLabs positions itself on the forefront of hybrid AI agents. Read more.
5 latest AI-powered tools from around the online
arXiv is a free online library where researchers share pre-publication papers.
Your feedback is useful. Reply to this email and tell us how you think that we could add more value to this text.
Inquisitive about reaching smart readers such as you? To turn into an AI Breakfast sponsor, reply to this email or DM us on 𝕏!
