标签：#AI

I have a problem with my social media feed: there are some people whose content I find really valuable, but on the other hand, there is too much trivial content, so I waste too much time watching low-value videos... BAD DEAL!!! So, let’s solve this issue using AI automation. We can solve this issue by using Telegram as our own social media platform and using automation to make our own feed, far from the algorithms of social media platforms. Choose the people whose content you find really valuable. Use AI automation to monitor them whenever they publish a new video. Once a new video is published, AI sends it to a Telegram bot where you can download it. The person who gave me this idea is Abo Ziad (the Egyptian version of Ali Abdaal), and he is one of my favorite YouTubers. Personally, I think you can add some entertaining channels to make it more balanced. I’ll try this solution for one month, and if you’re interested, I’ll share my experience with you, whether it succeeds or fails. In the end, it’s up to you. Tell us in the comments if you found the general idea useful or not, and whether AI automation is the best solution for this issue.

On the latest episode of Equity podcast, we discuss why not everyone is buying Zuckerberg’s vision.

"Breakups are… tough." It's the opening lines of an interlude towards the end of DOOMED! Called "It Happens in Every Universe." It's also basically the thesis of the entire record. It's no grand revelation, but it's a well-trodden subject that Open Mike Eagle manages to mine for artistic gold. Eagle's subject matter is usually personal, […]

This is a submission for Weekend Challenge: Dog Days Edition Target categories: Best use of Snowflake + Best use of Solana. Ten days from now International Dog Day is August 26th. The date is not arbitrary and it is not a marketing pick. Colleen Paige founded it in 2004 and chose the 26th because that is the day her family brought her first dog home from a shelter , when she was ten years old. So the holiday is not really about dogs. It is about the gap between a dog sitting in a shelter and a dog sitting in a house, and about how many animals never cross it. That gap is the whole reason this project exists, and it is why the last tab of this dashboard is the quietest one. I did not start at the shelter. I started at a limp nobody saw. But every path I traced ran the same direction, and it always ended in the same room. This is a build post, so most of what follows is SQL. But I want to be honest about which end of the problem I was standing at when I wrote it. The problem, stated plainly Because a hackathon post should be able to say this in four rows before it earns the right to show you any SQL: The question In one line The issue A dog in pain is built to hide it, so the first human-readable sign of a chronic problem arrives months late - often at the point the relationship, not just the joint, has broken down. Why software has not fixed it Every consumer tracker compresses a day of movement into a scalar - steps, active minutes, a sleep score - and then thresholds it. The clinical signal is not in the magnitude. It is in the ordering , and averaging is precisely the operation that deletes ordering. What TELLTAIL tries Stop thresholding. Make the detector a regular expression over rows - MATCH_RECOGNIZE - so a differential diagnosis stays a sequence all the way down to the individual second that satisfied it. How you know it is not a demo Every finding is explainable to the second, the accuracy is printed in 44px type on the dashboard including the parts that are

When I decided that I wanted to seriously start learning Artificial Intelligence, I quickly realized that one of the hardest parts wasn't finding resources. It was figuring out where to start. There are countless courses, YouTube playlists, roadmaps, tools, frameworks, and technologies to learn. Every time I looked at what other people were doing, I felt like there was something else I should be learning. So instead of trying to learn everything at once, I decided to create a roadmap for myself. This isn't a roadmap written by an AI expert or someone who has already mastered everything. It's simply the roadmap I'm following as a B.Tech Computer Science (Artificial Intelligence) student who is still learning. And I'm sharing it publicly because I want to document what works, what doesn't, and how my understanding changes along the way. Why I Decided to Learn AI Seriously I'm studying Computer Science with Artificial Intelligence, so AI has naturally become one of the areas I want to explore deeply. But for a long time, I didn't really know how to approach it. I knew that AI was important. I knew that Machine Learning, Deep Learning, and other AI technologies were becoming increasingly relevant. But knowing that something is important and actually learning it are two completely different things. After spending a lot of my first and second year without doing as much as I wanted, I realized that I couldn't keep waiting for the "right time" to begin. I had to start somewhere. So I decided to stop worrying about learning everything at once and focus on building my foundation first. Step 1: Strengthening My Programming Foundation Before jumping deeply into Machine Learning, I want to become more comfortable with programming. Python is one of the main languages I'm using for my AI journey because of how widely it is used in data science and Machine Learning. Alongside Python, I'm also learning C++ for Data Structures and Algorithms and working with Java for my college studi

💬 Following up on the story about the release of Qwen3.8-Max , I finally tried it on real-world tasks. Specifically, for building AI consultants for text channels (messengers) in my favorite programming language — Go . Spoiler: it’s really good, especially for such a low price per 1M tokens ! 😍 As a result, I built 2 demo Telegram bots, where GPT-4.1-mini acts as the brains 👇 1️⃣ A bot for qualifying a customer and booking a car repair appointment , which asks for details about the vehicle and the issue, answers questions about service pricing, and schedules a convenient visit time. 2️⃣ A bot for calculating kitchen pricing for furniture companies , which уточняет kitchen parameters through guiding questions, calculates the cost, sends the final estimate, and books the client at the company office for a detailed design session. Before implementation, of course, I wrote a detailed spec for each of these bots and connected MCP Context7. I also had to make 1–2 corrective prompts for code style and some business-logic details... but otherwise, Qwen3.8-Max worked fully autonomously in the engineering loop (questioning itself at every stage and adjusting its own reasoning and code). Token usage (input + output) totaled ~12.8 million , across about 400 API requests to the Chinese model. That’s seriously impressive! For comparison, I ran the same task through DeepSeek V4 Flash Latest: with similar output results, it used over 15 million tokens. By the way, the whole development process was done in the next-gen AI IDE Kodik , by our local guys — ArchiTech AI . Highly recommend downloading and trying it. Not an Ad! I’ve been using it for over a month now, and it’s truly a very high-quality product, especially in the era of account bans from Anthropic and OpenAI 😏 ...and soon, a local model called Qwen3.8-27b is also expected to drop, which Alibaba has promised to release any day now... that’s definitely something that can make the big AI model vendors nervous! 😉 And if you ne

Dario Amodei is pushing back against the idea that he's been painting an overly pessimistic picture of AI.

Welcome back to TechCrunch Mobility — your central hub for news and insights on the future of transportation.

Your agent calls get_weather(city="London") . The provider ships a new model version. Now it calls get_weather(location="London, UK") , your downstream parser breaks, and nothing in CI told you. I built a small library for exactly this failure: pip install toolcontract GitHub: https://github.com/Divyansh2202/toolcontract PyPI: https://pypi.org/project/toolcontract/ You pin a golden set of expected tool calls as a contract, re-run them against the live model, and get pass / fail / inconclusive with a diff showing what changed. It is not an eval framework. promptfoo, DeepEval and the rest score whether an output is good — semantic quality, usually judged by another model. toolcontract asks a narrower, cheaper question: is the tool call structurally the same as the one I pinned? Same tool, same argument shape, same trajectory. That is a regression test, not an eval, and it is the question that matters when a provider bumps a version under you. Details: pass / fail / INCONCLUSIVE — anything the structural comparators cannot resolve is never silently turned into a pass or a fail trajectory matching: strict, unordered, subset, superset optional argument support, so you can assert a field must stay absent works with OpenAI, Anthropic, anything OpenAI-compatible, or via LiteLLM thin pytest plugin, and a CLI that produces the same verdicts without pytest MIT Happy to hear where this breaks. It's v0.1.

With all the hype in the past several years around AI (or more specifically GenAI), I'm not afraid to say – I'm an AI skeptic. It doesn't mean that I don't believe that some day AI may have a huge impact on human beings' lives, but at the moment, all I can see is irrational hype. In my background, I came from infra-security; I am not a developer, nor do I consider myself an AI expert. I am a cloud architect, meaning I'm looking at proposed architectures, seeing how they suit business requirements, and whether they are deployed in a secure, resilient, and perhaps cost-effective way. I don't see value in adding AI to every design, just for the sake of saying "our application now includes AI". I've been watching the industry since 2023 go nuts. Suddenly, everyone is eager to add AI capabilities, chasing some unexplained FOMO before the machines replace our jobs. I'm not against the use of AI. As a matter of fact, I've been using Grammarly for many years (since, for most of us, English is not our first language). In the past several years, I've been using chatbots such as ChatGPT, Perplexity, and recently Gemini daily, asking questions about various topics and aspects of my life. From asking the bot to provide me an answer about a specific character in a favorite TV show, to "how do I resolve an alert shown on my car's dashboard," and up to "summarize this blog post for my newsletter". It's great that I can ask Gemini to create me a LinkedIn post based on an article I just read, add some emojis and hashtags, and at the end create me a cover image for the post. For a probabilistic system, this is great. I am expecting the system to be creative and produce me attractive results, sometimes even funny images. For a home consumer, this is great, but far from been ground breaking technology. I truly believe that the "big money" will come from enterprises paying a lot of money for AI-based solutions, once the industry can actually make something good from a non-deterministic s

When Claude Opus 5 and GPT-5.6 arrived, I expected my coding agents to become noticeably better. Instead, some of my workflows felt worse. The agents seemed more eager, less predictable, and occasionally “dumber” than before. Naturally, I blamed the new models. Very scientific of me. Maybe it was a skill issue Then I watched Andrej Karpathy’s interview on the No Priors podcast. One idea stuck with me: when an agent fails, the capability may already exist. The problem could be how we instruct it, what memory we provide, or how we arrange the workflow. That made me question something I had mostly ignored: Were my custom skills still compatible with the newer models? I gave my agent this prompt: Can you audit our custom skills against the current models? Flag stale prompts, conflicting instructions, outdated assumptions, and anything that should be simplified or removed. Then test each skill on a representative task and propose the smallest updates needed. The audit found instructions written around the behavior of older models. Some were redundant. Some were no longer necessary. Others pushed the newer models too hard and caused them to overdo tasks. After cleaning those up and testing the skills again, the results felt noticeably better. The official guidance supports this This was not proof that every disappointing result is caused by an outdated prompt. Models can still regress, behave inconsistently, or introduce genuine breaking changes. But both Anthropic and OpenAI recommend recalibrating instructions during model migrations. Anthropic’s Claude Opus 5 documentation says the model now verifies its work without being told. It specifically recommends removing verification instructions carried over from earlier models because they can cause over-verification. OpenAI’s GPT-5.6 guidance recommends removing repeated instructions, simplifying tool descriptions, and running the same evaluations after each change. In OpenAI’s internal coding-agent evaluations, leaner sys

DeepSeek Code is an open-source CLI built specifically for DeepSeek — one of the most cost-effective AI models on the market! It was heavily inspired by tools like Claude Code and OpenAI's Codex for most of its feature set. Since it is fully open-source, you are more than welcome to open Pull Requests, report bugs, or submit issue suggestions! Installation You can install it globally via npm or bun: `` Using npm npm install --global @hermenics/deepseek-code Using bun bun add -g @hermenics/deepseek-code `` Check out the repository, star the project, or contribute: 👉 GitHub: https://github.com/Hermenics/deepseek-code Feel free to test it out and leave your feedback below! PRs and Issues are highly appreciated. ai #cli #showdev #typescipt

On 11 August the GoFullPage extension disappeared from the Chrome Web Store and got disabled in Chromium browsers. Eleven million users, one Tuesday. It was not a hack. The developers say it was a copyright dispute over a design element, that it was "definitively not a security issue", and that they are working with Google on getting it back. Chrome shows the same "might be unsafe" string for every kind of Web Store policy breach, so the warning read far worse than the cause. Two things came out of it. A short list of fixes, and a longer thought about where our tools live. Get working again in a minute Re-enable it. Open chrome://extensions . If Chrome disabled the extension rather than deleting it, the toggle is still there. Use Edge. GoFullPage was never removed from the Edge add-ons store. Install the beta. The team published a separate build at ID kehafhfdnkhdgbnpeofmhmbibmpnjaof , and it can sit alongside the original. That is the practical answer. The more interesting one is that most of us never needed the extension. Five ways to capture a full page with no extension at all 1. Chrome DevTools, no code Open DevTools, press Cmd/Ctrl + Shift + P , type screenshot , choose Capture full size screenshot . This has been in Chrome for years and most people have never found it. It handles scroll-height pages properly and drops a PNG in your downloads. 2. Firefox, even shorter In the Firefox console: :screenshot --fullpage Add --dpr 2 for a retina-density capture, or --clipboard to skip the file. 3. Chrome DevTools Protocol, if you want it scripted The thing the extension was wrapping is one CDP call: await client . send ( ' Page.captureScreenshot ' , { format : ' png ' , captureBeyondViewport : true , }); captureBeyondViewport is the flag that does the work. Everything else in a full-page screenshot tool is UI around it. 4. Playwright import { chromium } from ' playwright ' ; const browser = await chromium . launch (); const page = await browser . newPage ({ viewport

ChatGPT's desktop app on macOS has a new feature called Computer History that turns your actions into training data, learning how you work, suggesting automations, and even picking up tasks you left half done. It uses your activity to build a timeline that ChatGPT and Codex can reference when you make a request. The feature […]

Let's get something out of the way first. Having data is good. Having a database full of reviews, commits, and org activity sitting there quietly, untouched, unread, never once glanced at by a human being with a coffee and an opinion? That's not "having data." That's a very expensive data graveyard. At LiveReview , we build what we call a Blast-Radius Aware AI Code Review for Business-Critical Systems . Which is a fancy way of saying: we review your code, we figure out how bad it would be if a change goes wrong, and we don't shut up about it until someone fixes it. Along the way we accumulate a review data: who reviewed, how much, how fast, how often, which repos are on fire. And for a while, that pile just sat there. Engineering leaders would ask "is adoption increasing?" and get back a vibe, not an answer. So we built Livi , a chat bot that answers real questions about that data with real charts, not paragraphs of hedging. This post technically about how Livi draws those charts. Specifically: why we never let the LLM touch a pixel, how the same chart definition ends up as both a live interactive graph in your browser and a flat PNG in a Slack thread, and why teaching a language model to pick the right chart shape is a surprisingly deep rabbit hole. The core decision: don't ask the LLM to draw, ask it to describe The tempting, wrong idea is: "let's have the LLM generate an image." Please don't. Image-generating models are a different beast entirely, and even if you got one to draw a bar chart, you'd have no way to verify the numbers on it are real. You'd be trusting a model that hallucinates plausible-sounding review counts to also render them faithfully into pixels. That's not a chart, that's chart-shaped fan fiction. The actually good idea, and the one every serious LLM-charting integration eventually converges on, is: the LLM writes Vega-Lite , a JSON grammar for describing charts declaratively. You don't say "draw a blue bar going up." You say: { "mark" : "bar"

This tutorial walks through building an MCP server in Rust with rmcp , the official Model Context Protocol Rust SDK. The example is a real one: a devops agent that manages AWS EC2 G5g instances — Graviton2 boxes with NVIDIA T4G GPUs — serving Gemma 4 under vLLM. It launches instances, drives them over SSM, and health-checks the model. There's an existing Python version, so at the end we can put the two side by side. Follow along and you'll have a working, registerable MCP server. 🦀 Why Rust for this? Worth answering properly, because the weak version of the argument is easy to make and easy to demolish — and the real one is better anyway. Start with what it isn't: these tools are I/O bound. Every one is an AWS API call — describe_instances , send_command , polling SSM — so 100–500 ms of network per call. The caller's language contributes nothing measurable there. Anyone selling you a Rust rewrite on raw speed for this workload is selling something. Three claims that don't hold, so nobody has to make them in the comments: Claim Why it fails "462 ms startup is slow" stdio servers spawn once per session , not per call "Rust is faster" the work is network round-trips to AWS "smaller supply chain" 241 crates vs 34 Python packages — it's worse What actually justifies it, for this codebase: 1. It's a fleet, not a server. This monorepo has 16 rigs , each with its own MCP server. That changes the units: All loaded together 🐍 Python 🦀 Rust Resident memory 16 × 83 MB ≈ 1.33 GB 16 × 12 MB ≈ 192 MB Session startup 16 × 462 ms ≈ 7.4 s 16 × 2.5 ms ≈ 40 ms A gigabyte of resident Python to expose sixteen tool lists is a real cost. 2. No shared interpreter. These rigs install system-wide — no virtualenvs, by policy — so all sixteen share one Python. Sixteen servers with independently drifting boto3 and mcp pins in one interpreter is a standing conflict risk. A static binary has no such coupling; each rig pins whatever it likes in its own Cargo.lock . 3. The schema can't drift from th

This tutorial walks through building an MCP server in Rust with rmcp , the official Model Context Protocol Rust SDK. The example is a real one: a devops agent that manages AWS EC2 G5g instances — Graviton2 boxes with NVIDIA T4G GPUs — serving Gemma 4 under vLLM. It launches instances, drives them over SSM, and health-checks the model. There's an existing Python version, so at the end we can put the two side by side. Follow along and you'll have a working, registerable MCP server. 🦀 Why Rust for this? Worth answering properly, because the weak version of the argument is easy to make and easy to demolish — and the real one is better anyway. Start with what it isn't: these tools are I/O bound. Every one is an AWS API call — describe_instances , send_command , polling SSM — so 100–500 ms of network per call. The caller's language contributes nothing measurable there. Anyone selling you a Rust rewrite on raw speed for this workload is selling something. Three claims that don't hold, so nobody has to make them in the comments: Claim Why it fails "462 ms startup is slow" stdio servers spawn once per session , not per call "Rust is faster" the work is network round-trips to AWS "smaller supply chain" 241 crates vs 34 Python packages — it's worse What actually justifies it, for this codebase: 1. It's a fleet, not a server. This monorepo has 16 rigs , each with its own MCP server. That changes the units: All loaded together 🐍 Python 🦀 Rust Resident memory 16 × 83 MB ≈ 1.33 GB 16 × 12 MB ≈ 192 MB Session startup 16 × 462 ms ≈ 7.4 s 16 × 2.5 ms ≈ 40 ms A gigabyte of resident Python to expose sixteen tool lists is a real cost. 2. No shared interpreter. These rigs install system-wide — no virtualenvs, by policy — so all sixteen share one Python. Sixteen servers with independently drifting boto3 and mcp pins in one interpreter is a standing conflict risk. A static binary has no such coupling; each rig pins whatever it likes in its own Cargo.lock . 3. The schema can't drift from th

LLMs can be surprisingly useful for company research. But I kept running into a strange split: parts of the reasoning were useful, while the financial facts underneath them were much harder to trust. A model could identify an accounting risk in one paragraph, then mix fiscal periods, accounting scopes, or currencies in the next. Missing values might quietly become zeros. A deterministic calculation could be performed probabilistically. A citation could point to a real filing without actually supporting the claim. Those are different failure modes, and treating all of them as one giant prompting problem did not feel like a reliable architecture. So I started building OpenThesis , an Apache-2.0 desktop system for evidence-first, AI-assisted company research. The project is not a stock picker or a trading bot. The idea is simpler: use ordinary software for work that should be deterministic, and give the LLM a bounded evidence set for the reasoning work where it can actually help. The monolithic prompt is doing too many jobs A common company-research workflow looks roughly like this: company question ↓ LLM ↓ answer That single model call is implicitly responsible for remembering reported values, selecting the right fiscal period, recognizing the accounting scope, finding sources, performing calculations, comparing scenarios, identifying risks, and writing a conclusion. Some of those tasks are probabilistic by nature. Others are not. Qualitative reasoning, connecting evidence, forming scenarios, and challenging an assumption are reasonable uses of a language model. Remembering an exact reported value, deciding whether a value is missing, and calculating a margin or valuation are poor places to accept probabilistic behavior. My design rule became: Deterministic work should stay deterministic. Use LLMs for reasoning, not as the database and calculator underneath the reasoning. Evidence before reasoning OpenThesis starts from official filings rather than from model memory o

Built 2026-08-15 against us.amazon.nova-lite-v1:0 via the Bedrock Converse API. FastAPI on Python 3.13, deployed to an Amazon Lightsail container service ( nano , scale 1) in us-east-1 . Scored against the live deployment, not localhost: 20/20 on the fixture set, median 880 ms per scan. Live: Dog or Not: Lite · Source: github.com/xbill9/dog-or-not-lite · Built for the AWS Weekend Challenge: Build a Creative App . TL;DR Make the model fill in a schema instead of writing a sentence. The Converse API's toolConfig plus toolChoice forces a named function call, so is_dog arrives as a boolean because it was declared as one. Every image comes back in the same shape — including the ambiguous ones, which is exactly where free-text output gets creative and a string-matching parser gets it wrong. The app is a webcam scanner that tells you whether the thing you are holding up is a dog. One HTML page, one POST /api/scan , one model call, no build step, no framework. The whole backend is 285 lines. Three AWS specifics are worth the price of admission: Lightsail container services have no IAM task role. There is nothing to attach a policy to, so the container needs a real access key as an environment variable. The mitigation is scope, not secrecy. A cross-region inference profile is authorized against every region it routes to. With the policy pinned to us-east-1 , a call made to us-east-1 was denied naming us-west-2 . Measured, not inferred. --platform linux/amd64 is not optional. An arm64 image builds, pushes and deploys cleanly, then crash-loops with an exec format error that never mentions architecture. And a mock mode that answers every scan locally is what made the frontend free to build — no credentials, no model access, no bill. 1. The shape: one route, one call The classification rule is the only opinionated part. is_dog is true only for a living domestic dog: a wolf is not a dog , nor is a coyote, fox, plush toy, bronze statue, cartoon, or person in a costume. That is a c

This is the data appendix for Posts 1–4 . The narrative and takeaways live in the main posts. This page is pure measurement. Eval set: 22 questions (P1) → expanded to 31 questions (P2 onward) Corpus evolution: P1/P2: ~3,700 docs (raw threads + code chunks) P3/P4: 16,315 docs (distilled threads + bursts + code) Quick comparison (same 31-question set) Metric Vector P2 Hybrid P2 Vector P3 Hybrid P3 Hybrid + Rerank (P4) recall@1 0.68 0.61 0.52 0.39 0.87 recall@3 0.84 0.65 0.71 0.65 0.94 recall@10 0.90 0.90 0.81 0.94 0.94 MRR 0.77 0.67 0.63 0.57 0.90 Takeaway: Hybrid alone never beat pure vector on this corpus. Hybrid + LLM rerank is the first clear win. P1 — Naive vector baseline Corpus: 3,000 raw issue threads + 687 code chunks Embeddings: BGE-M3 (1024d), max_seq_length=1024, HNSW cosine Numbers (22 questions) Metric Score recall@10 1.00 (22/22) recall@3 0.95 recall@1 0.77 (17/22) Main k=1 misses Exact error pastes ( TypeError: Object of type int64... , AttributeError: 'Depends'... ) — ranked 4–5 instead of 1 jsonable_encoder code chunk outranked by issues about the function API key header implementation (code vs similar issues) Paraphrase questions (dependency injection outside routes, custom 404) Pattern: Dense search is strong on recall@10 but weak when the query has a sharp lexical signal. Ops notes Ingest wall time ~40 min (GitHub API is the bottleneck) BGE-M3 OOM on Apple Silicon fixed by capping max_seq_length=1024 Python 3.13 + uv editable install issue fixed by pinning 3.12 P2 — Hybrid (vector + FTS + RRF) Corpus: Same size as P1, with better comment pagination and symbol-based code IDs Eval set: Expanded to 31 questions (added exact error pastes + rare identifiers) Numbers Metric Vector FTS Hybrid recall@1 0.68 0.42 0.61 recall@3 0.84 0.48 0.65 recall@10 0.90 0.65 0.90 MRR 0.77 0.47 0.67 Headline: Hybrid is not a strict win over vector-only. Where hybrid helped Exact error pastes (e.g. TypeError: int64 is not JSON serializable ) → moved from rank 5 → 1 Near-d