标签：#t

DeepSeek has previewed its V4 model family, led by a 1.6 trillion-parameter flagship, and made a one-million-token context window the default across all its services. The weights are downloadable and self-hostable, putting frontier-scale long context in reach of smaller labs and individuals without per-token payment to a closed provider. Key facts What: DeepSeek previewed two free-to-download V4 models that can read a million tokens at once, no longer as a premium add-on but as the standard setting. When: 2026-06-29 Primary source: read the source A large language model has no persistent memory. Each time it answers, it re-reads everything in front of it — your question, the conversation so far, any documents you pasted — and that pile of text is the context. The context window is the hard ceiling on how much it can hold at once. For years that ceiling was a few thousand words, then tens of thousands. Pushing it to a million has been possible but expensive, usually sold as a special, pricey tier. DeepSeek's move is to make a million the everyday default. The family comes in two sizes. V4-Pro is the big one — 1.6 trillion parameters in total, but only about 49 billion of them switch on for any given word. That design is called a mixture of experts : instead of running the entire brain for every token, the model routes each piece of text to a small relevant subset of specialists, so it stays affordable to run despite its enormous size. V4-Flash is the smaller, cheaper, faster sibling, meant for everyday chat and quick edits, and DeepSeek says it keeps up with Pro on simpler agent tasks. Making a million-token window affordable comes down to how the model handles its KV cache — the running set of notes it stores about every previous word, which grows steadily the longer the conversation gets. At a million tokens those notes become a mountain of memory, and the model normally has to consult every note for every new word it writes. DeepSeek's approach, which they call sp

A note before we start: this is about the machine, not the money. I'm not going to show you returns, positions, or a single "this strategy made X%." Partly because that's a regulatory minefield, and partly because the returns aren't the interesting part — the engineering is. If you came for a get-rich screenshot, this isn't that. If you came to see how one person ships production infrastructure with an AI, pull up a chair. The thing I built Over the last few months I built, with an AI coding agent as my pair-programmer, a fleet of ~35 automated trading bots. They run across five equity markets plus crypto. Each one is a long-running service. They share a single database, post to a live dashboard, fire alerts to my phone, and — the part that took the longest — they're built to survive restarts, reconcile against reality, and refuse to do anything stupid. I'm one person. I am not a team. The "team" is me plus an AI in a terminal, working the way you'd work with a very fast, very literal junior engineer who never gets tired and occasionally needs to be talked out of a bad idea. Here's how it's put together, and the handful of lessons that cost me the most to learn. The architecture, in one breath One Postgres database is the brain — every trade, signal, and piece of state lives there. Around it sit ~35 containerized bots, each isolated (its own tables, its own config, its own identity), orchestrated with Docker Compose. A Streamlit dashboard reads the database and renders the whole fleet — open positions, P&L curves, health. A notification layer pushes Telegram alerts on every meaningful event. Schema changes go through migrations so a new bot is never born with a stale database shape. Each bot is the same skeleton wearing a different hat: a signal module (the strategy logic), a trader that turns signals into orders, a storage layer that persists everything, a runner loop on a schedule. Strategies are swappable. The infra underneath them is identical. That sameness is

Far-UVC could help rid future cars of odors.

The AI neocloud provider, which specializes in hosting open source models, last raised at a $3.3 billion valuation in early 2025.

Tariffs and anti-China policies killed this little Volvo in the United States.

The nine-year-old scooter and bike-share company has said it needs the funds to help pay down around $1 billion in liabilities.

We started using Jira to manage our internal development workflow. At first it worked fine, but once we outgrew the free tier, the cost became hard to justify. At $15 per user per month, we were suddenly looking at a bill that did not match how we actually used the product. What we Built We created WannaTrack, a lightweight project management tool designed for small dev teams that do not need enterprise complexity. The goal was not to recreate Jira. It was to remove everything we did not use. Key ideas : minimal agile board with no clutter or heavy configuration simple issue tracking flow fast interface for daily development work minimal setup and no onboarding overhead Migration from Jira One of the biggest concerns was switching tools without breaking our workflow. So we built a Jira import tool that lets you migrate existing tickets into WannaTrack without manual effort. This allowed us to switch internally without downtime. Where it is now We now use WannaTrack daily for our own development workflow and are opening it up to other teams who feel the same pain with traditional tools. If you are a small dev team, indie hacker, or startup looking for a simpler issue tracker without overhead, you can check it out here: https://wannatrack.com

If you've ever tried to build a simple command/telemetry protocol between a PC and a fleet of SDR receivers, sensors, or embedded devices, you know the usual options aren't great: Roll your own binary format — fast, but you end up writing and maintaining custom serialization code for every device type, and debugging mismatched structs across machines is painful. Protobuf / FlatBuffers — robust, but require you to define your message layout in a schema file upfront, run a code generator as part of your build, and commit to a fixed structure. Adding a new device type or a new parameter means editing the schema, regenerating, recompiling everything. JSON over the wire — easy to debug, but heavy for anything real-time or bandwidth-constrained. I ran into this while working on a multi-SDR receiver system and ended up writing MessageFrame — a small C++17 library that lets you build structured messages dynamically, without any schema files or code generation. The basic idea Instead of defining a struct for each device type, you address each parameter with two strings — a device name and a parameter name — and the library handles the rest: // One message, multiple devices, assembled at runtime msgframe :: MessageFrame msg ( MSG_TELEMETRY , TYPE_PERIODIC , src = 1 , tgt = 2 ); msg . add ( "sdr_1" , "rx_gain" , VALUE ( 30.0 )); msg . add ( "sdr_1" , "center_freq" , VALUE ( 915'000'000.0 )); msg . add ( "sdr_1" , "sample_rate" , VALUE ( 2'000'000.0 )); msg . add ( "sdr_2" , "rx_gain" , VALUE ( 25.0 )); msg . add ( "sdr_2" , "lock_status" , VALUE ( true )); msg . add ( "psu_1" , "voltage" , VALUE ( 12.04 )); msg . add ( "psu_1" , "temp_c" , VALUE ( 47.3 )); // Attach raw IQ data alongside the parameters std :: vector < uint8_t > iq_buffer = { 0x01 , 0x02 , 0x03 , 0x04 }; msg . add_attachment ( "raw_iq" , std :: move ( iq_buffer )); // Serialize into a buffer, send over whatever transport you use std :: vector < uint8_t > out ; msg . serialize ( out ); send_udp ( out . data (),

As of May 2026, more than 80% of the code Anthropic ships is written by Claude, not by its human engineers. The company disclosed the figure in an essay called When AI builds itself , with coverage from Tom's Hardware and VentureBeat . Key facts What: Anthropic says more than 80 percent of the code it ships is now written by its own model, Claude, and the more interesting numbers are about judgment. When: 2026-06-23 Primary source: read the source Two years ago this share sat in the low single digits. The shift accelerated after Anthropic released Claude Code , a tool that lets the model read an entire codebase, make changes, run tests, and fix what breaks without human help. The human role has flipped: engineers used to author the code while the machine assisted; now the machine authors the code and engineers review, approve, reject, and steer. Anthropic reports its typical engineer ships roughly eight times as much code per quarter as a few years ago — not because people type faster, but because they spend their day reviewing the model's output instead of writing from scratch. Think of it as a newsroom where a tireless junior writer drafts every article and senior editors only sign off. Volume goes way up. But the 80% figure is less impressive than it sounds: a draft that a human must check, fix, and approve is not the same as a writer you can leave unsupervised. Most of those lines still pass through a person. On its own, this number measures effort the machine saves, not work it can be trusted to do without oversight. The results buried deeper in the essay matter more, because they concern taste rather than volume. Anthropic ran a recurring test where the model chooses the best next step in a research project, then compared its choices against its own scientists. Late last year the model was roughly a coin flip against the humans. By spring 2026, an unreleased internal model was picking the better direction clearly more often than its own researchers. Choosing w

Are you worried your AI chatbot is trying to build a bomb or leak personal information about you? There’s a website for that.

Cloudflare is giving AI companies until September 15 to separate web crawlers used for search from those used for AI training and agents, or risk being blocked by default on many publisher sites.

After a lengthy legal dispute, Krafton has settled with its subsidiary Unknown Worlds Entertainment, which is developing Subnautica 2, and will pay bonuses to the studio's staff, Bloomberg reports. The dispute began last year after Krafton pushed out Unknown Worlds' cofounders, Charlie Cleveland and Max McGuire, and its CEO, Ted Gill, ahead of a potential […]

Microsoft will likely soon follow Sony and stop the production of physical discs for Xbox games. But instead of leaving physical discs behind entirely, sources familiar with Microsoft's plans tell me the company has quietly been working on a disc-to-digital feature that will allow Xbox owners to digitize their existing physical game collections. Xbox employees […]

Is PlayStation about to make the same mistake Microsoft made with Xbox One?

Honda wants in on the lucrative energy storage market. This week it began producing batteries destined for data centers, not driveways.

"You gods don't speak in ways we understand."

The autonomous vehicle space is starting to feel like a repeat of the 2016 hype cycle. Travis Kalanick is back building a robotics company, and the talent wars and capital are heating up the same way they did the first time around. The money’s flowing back, and it’s the people who lived through that first […]

US lifts curbs on Anthropic’s advanced Fable and Mythos models.

Kobo recently raised the price of its Libra Colour e-reader to $259.99, but today’s deal effectively erases that hike. The company, Best Buy, and Target, are all selling it for its old $229.99 price as part of broader sales, making now one of the best times to grab my favorite alternative to Amazon’s Kindle. If […]

Penalty kicks are already proving critical to big wins at this year’s World Cup. But the advantage in penalty kicks has more to do with psychological effects than who kicks first.