AI News

Install the LLM gateway by asking the agent: Keystone plugin for Claude Code / Habr

May 4, 2016 | Habr AI


Install the LLM gateway by asking the agent: Keystone plugin for Claude Code / Habr
✔ Successfully added marketplace: ai-process (declared in user settings) Installing plugin "ktms@ai-process"...✔ Successfully installed plugin: ktms@ai-process (scope: user)Skills call CLI ktms, it needs to be placed in PATH from the release archive or through go install. The plugin version matches the Keystone version.SkillWhat it does/ktms:assessLooks at the system and tells where Keystone is installed and where to start/ktms:deployCompose or Helm, secre

Helping the little one become strong: Local Qwen vs. cloud flagship models with 16 GB of video memory / Habr

September 8, 2018 | Habr AI


Helping the little one become strong: Local Qwen vs. cloud flagship models with 16 GB of video memory / Habr
The cat has been optimized.

Manchu language corpus. How we taught Tesseract to read Manchu / Habr

April 28, 2017 | Habr AI


Manchu language corpus. How we taught Tesseract to read Manchu / Habr
By the way, in some consoles, the macron in the second variant is noticeably shifted. This is the same rendering issue that forced us to adjust character frames in step 3. In our scripts, such normalization is performed for each text before rendering. After this, each letter has exactly one representation. The model sees the same codes for identical letters and doesn't learn anything extra.Step 2. Drawing strings with different fontsPrinted publications ar

How we connected the AI bot to MAX: Bot API, common bridge with Telegram and deep links / Habr

March 30, 2023 | Habr AI


How we connected the AI bot to MAX: Bot API, common bridge with Telegram and deep links / Habr
Bridge — a separate service in the same docker compose assembly as the rest of the core. It knows nothing about agents. It can do three things: fetch an event from MAX, convert it into an internal message, and send a response back.MAX Bot API: what is neededThe bot works on the official MAX Bot API (platform-api.max.ru). The bot is registered through the MAX partner account.Events are fetched using long-polling, without webhooks. For one bot it's simpler:

My AI agents got an office: Lumio 1.4 — Claude Code, Codex, Qwen, and Cline in the Mac / Habr lineup

February 1, 2020 | Habr AI


My AI agents got an office: Lumio 1.4 — Claude Code, Codex, Qwen, and Cline in the Mac / Habr lineup
Everything is visible from the menu bar Opening the menu bar reveals an island with a list of sessions. Each has a clear status: "writing code," "reading files," "waiting for you," "complete." If an agent is requesting permission for a command, the "Allow" and "Deny" buttons are right on the card: one click, and the agent proceeds, while you stay where you were. Questions with options can also be answered from the menu bar. Lumio understands eleven agents:

Lightweight LLMs on CPU: What real tasks can be done on an affordable VPS / Habr

September 3, 2020 | Habr AI


Lightweight LLMs on CPU: What real tasks can be done on an affordable VPS / Habr
models from Hugging Face and most of the registry are in quantization Q4_K_M. Larger Russian models, such as GigaChat 3.1 Lightning, T-lite, and YandexGPT-5-Lite, didn't fit into 4 GB and are waiting for tests on a larger server. Before the measurements, I had to spend some time with the server and do four things: Swap 4 GB with vm.swappiness=10. Without it, the kernel would terminate 4B models on long texts due to memory shortage, but with swap they work,

AI is not a magic wand. How to work with a specification agent and not fall into a bug loop / Habr

October 25, 2017 | Habr AI


AI is not a magic wand. How to work with a specification agent and not fall into a bug loop / Habr
The feature goes through a cycle of three commands:/opsx:propose add-partial-refund: the agent creates a change folder with proposal, specification deltas, design and tasks. On this step, you review the plan until no line of code has been written./opsx:apply: the agent implements tasks from tasks.md and marks completed ones./opsx:archive: the change goes into the archive, and deltas are merged into the main specifications in specs/.Requirements and plan li

Timnit Gebru, dismissed by Google, wins 'alternative Nobel'

October 22, 2025 | Olhar Digital IA


Timnit Gebru, dismissed by Google, wins 'alternative Nobel'
Everything about Google view more Everything about Artificial Intelligence view more The scientist Timnit Gebru won the Right Livelihood Award, known as the "Alternative Nobel". The award recognized her work on digitally based justice technologies.Former Google researcher Gebru was fired after refusing to remove conclusions from a study on biases in language models. The jury awarded her for questioning the concentration of power in technology. Timnit Gebru

Is AI making us stupid? What you need to understand - Olhar Digital

September 6, 2020 | Olhar Digital IA


Is AI making us stupid? What you need to understand - Olhar Digital
Everything about Artificial Intelligence view more Whether you like it or not, whether you use it or not, generative artificial intelligence "took off". The technology has deeply changed how we do things. And how we think about things.Currently, we are in the "AI agents" phase of this technology. At this level, AI does things for us, instead of just guiding us. Whether it does it well or not depends on factors such as: how much you paid to use the tool, ho

AI has 'a toll you'll have to pay later,' says psychologist - Olhar Digital

September 3, 2023 | Olhar Digital IA


AI has 'a toll you'll have to pay later,' says psychologist - Olhar Digital
Everything about Artificial Intelligence view more Artificial intelligence has changed quite a bit since the end of 2022, when ChatGPT was launched. Today, the technology probably speeds up one thing here and another there for you, whether in personal life or at work.Using the technology may seem harmless. But there's "a toll you'll pay later." That's what Cristiano Nabuco, psychologist and rector of Artmed School of Psychology (APSY), said in an interview

AI will have 4 levels of supervision after an agreement between big tech companies in the US

February 8, 2016 | Olhar Digital IA


AI will have 4 levels of supervision after an agreement between big tech companies in the US
Everything about United States view more Everything about Artificial Intelligence view more The leading artificial intelligence (AI) companies in the United States signed, on Tuesday (29), a voluntary code of conduct presented by Donald Trump's government. The agreement establishes commitments related to safety, innovation, and the development of AI technology.The document was signed during a meeting between Trump and executives of major sector companies a

Google: AI model advances in programming and cybersecurity

February 22, 2015 | Olhar Digital IA


Google: AI model advances in programming and cybersecurity
Everything about Google see more Everything about Artificial Intelligence see more The Google presented, on Wednesday (30), the Gemini 4 Argon, new artificial intelligence (AI) model that now leads the Gemini 4 line. According to the company, it is their most advanced model so far, developed to handle complex and long-duration tasks in areas such as programming, cybersecurity, and professional activities. The launch occurs after months of waiting for the s

Meta denies that Muse has read messages without authorization

January 21, 2018 | Olhar Digital IA


Meta denies that Muse has read messages without authorization
Everything about Artificial Intelligence view more Meta denied that its artificial intelligence (AI) agent Muse had accessed a user's private messages without authorization. The statement came after journalist Jason Aten, columnist at Inc., claimed that the system had accessed the message database of his Mac even after he did not grant permission for it.According to Aten, Muse began suggesting ideas for a column after he conversed with the co-host of his p

OpenAI accuses Chinese Moonshot of distilling its AIs - Olhar Digital

November 24, 2020 | Olhar Digital IA


OpenAI accuses Chinese Moonshot of distilling its AIs - Olhar Digital
Everything about Artificial Intelligence view more Everything about OpenAI view more OpenAI claims to have identified and interrupted a coordinated campaign of model distillation of artificial intelligence (AI) aimed at extracting protected reasoning capabilities from its systems. According to the company, a central part of the activity was attributed to individuals associated with the Chinese Moonshot AI, developer of the chatbot Kimi. The activity had st

Hackers use false identity to attack AI researchers

June 14, 2016 | Olhar Digital IA


Hackers use false identity to attack AI researchers
Identified as TA419, the group attempts to capture credentials of individuals associated with universities, research centers, defense companies, and law offices in the United States and Japan since at least 2025. In the latest campaign, the attackers used fake invitations for AI projects as bait. The fake emails proposed collaborations on artificial intelligence before directing the targets to fraudulent pages. Image: SuPatMaN/Shutterstock How the fraud wo

AI agents tried to invade the website of the Canadian government.

August 31, 2018 | Olhar Digital IA


AI agents tried to invade the website of the Canadian government.
Everything about Artificial Intelligence view more Everything about OpenAI view more Transluce, an artificial intelligence research company, revealed that AI agents attempted to access a Canadian government website. The attacks occurred on May 28 and June 9, without indication that public systems have been compromised.The target was the Library and Archives Canada portal. The activity drew attention after arquivo.pt, a Portuguese web archive, recorded 899

Caetano, Neymar e Tony Ramos: Artificial Intelligence transforms celebrities into weapons of disinformation and financial scams - Olhar Digital

February 8, 2022 | Olhar Digital IA


Caetano, Neymar e Tony Ramos: Artificial Intelligence transforms celebrities into weapons of disinformation and financial scams - Olhar Digital
Everything about Artificial Intelligence see more Fake videos using the voice and image of well-known figures have exploded on social media in recent weeks, especially with the arrival of elections. Recent cases show the power of artificial intelligence (AI) in this movement – it can invent political statements, alter audio, and create scenes so realistic they easily deceive someone scrolling on their phone screen.This tactic exploits the influence of popu

Rejected by Claude, the U.S. army turns to ChatGPT

March 19, 2026 | 01net IA


Rejected by Claude, the U.S. army turns to ChatGPT
Published on March 2, 2026 at 10:11 Commenter 0 𝕏 Print © Steven Abraham (Unsplash) Anthropic has chosen to say no to the Pentagon. Following an unprecedented showdown between an AI startup and the Trump administration, Claude will no longer be used by the armed forces. To replace Claude, the Ministry of Defense negotiated an agreement with OpenAI, the creator of ChatGPT. Anthropic says no to the U.S. military. The startup behind Claude decided to ban the

With its AI devices, Lenovo may have found the right formula

January 24, 2018 | 01net IA


With its AI devices, Lenovo may have found the right formula
Published on March 2, 2026 at 16:46

Apple could entrust the hosting of the new Siri to Google, by imposing its own privacy rules.

October 16, 2024 | 01net IA


Apple could entrust the hosting of the new Siri to Google, by imposing its own privacy rules.
Published on March 3, 2026 at 10:00 Comment 0 𝕏 Print © 01net The future Siri, which was supposed to start appearing in iOS 26.4, relies on Google's AI models. But it's not the only web giant's technology that the apple company needs. It should also entrust the hosting of the new assistant to Google servers, but configured according to Apple's data protection standards. In June 2024, Apple presented the new Siri, an AI-powered assistant capable of searchin