Table of Contents
- 1. CTV and Algorithm-Driven Viewing
- 2. Monitor Network Traffic Easily
- 3. Introducing HTTPie AI
- 4. A Lighter Touch Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X
- 5. Modern lsusb with New Features
- 6. The Startup M&A Market Fell 94% Year over Year - But One Segment is Thriving by @ttunguz
- 7. Object Storage Wins Big Data
- 8. AI Answers Banned on StackExchange
The world is changing rapidly, with it, the media we consume. With the emergence of TikTok, GitHub, HTTPie, AI, and the Core i9 CPU, the TV industry needs help to keep up. But what about the startup M&A market and databases? Is the market over? And what can tech learn from the pandemic? This article explores the power of scaling, the potential of generative AI, and the lessons businesses can learn from the toilet paper Magik.
CTV and Algorithm-Driven Viewing
TikTok openly talks about its ambition to conquer the "2 billion screens outside of mobile," which was emphasized at the Consumer Electronics Show in Las Vegas. This raises the question- Why isn't TV trying to get TikTok? Back in June, the app's incredible growth worried traditional media companies. People spend upwards of two hours daily on the app, which takes away time from traditional media companies. At CES, giant TVs abound, and the advertising wing demonstrated the power of streaming and the fast-growing CTV ad sector. Companies like Netflix and Disney are selling; everyone is doing FAST channels, including a new batch from Warner Discovery. TV looks set to be doing better than ever when combined with digital targeting. However, people spend their time watching short-form videos on TikTok. What is TV's answer to the TikTok threat other than a blind belief in CTV? Facebook and Google have released blatant TikTok ripoffs, which people use, and they may even make money from these products. Marketecture is a new publication offering in-depth interviews with leading vendors to help businesses decide which partners are the best fit.T
Added to Tech Innovation lists by @tobyoedwards
CTV and Algorithm-Driven Viewing
The TV industry seems to have no answer to TikTok
Short form, algorithm-driven viewing is here to stay, and CTV won't change that
Monitor Network Traffic Easily
Sniffnet is a cross-platform application to monitor your network traffic with ease. Git commands accept tag and branch names, so creating this branch may cause unexpected behavior. Use Git or checkout with SVN using the web URL to create the branch. T The application binary can be installed with Rust or installers available in the latest release. To run Sniffnet on Windows systems, add the SDK's /Lib or /Lib/x64 folder to the LIB environment variable. On Linux systems, install the libraries and header files for the libpcap library and set capabilities to inspect a network adapter. MacOS has all the required dependencies. Application layer protocols are inferred from the transport port numbers, and most errors occur due to missing cap dependencies. For Linux systems, libfontconfig may be necessary. If problems happen with the installers provided after installation, please contact an installer for the specific operating system.t
Added to tech tools lists by @mpicard
Monitor Network Traffic Easily
GitHub - GyulyVGC/sniffnet: Cross-platform application to monitor your network traffic with ease
Cross-platform application to monitor your network traffic with ease - GitHub - GyulyVGC/sniffnet: Cross-platform application to monitor your network traffic with ease
Introducing HTTPie AI
HTTPie AI is a new way to interact with application programming interfaces (APIs). It features an AI assistant, tabs for HTTP methods, a library context menu, cURL input mode, and more. The new version of HTTPie, version 20.22.0, is available for web and desktop and includes support for Postman and Insomnia imports and GraphQL improvements.T
Added to Tech History lists by @cbrown66
Introducing HTTPie AI
HTTPie AI – HTTPie blog
Today we’re launching HTTPie AI, a new way to interact with APIs. HTTPie AI assistant uses state-of-the-art artificial intelligence to help you become more productive.
A Lighter Touch Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X
The running joke on social media and Reddit is that CPUs have become increasingly popular. This is seen in the comparison between a flagship x86 CPU now, which requires a heatsink, and one in the past, which did not. Although the CPUs are not at the level of a nuclear reactor, as the jokes go, chipmakers still strive to make them more powerful by designing for higher TDPs and greater power consumption. This is seen in the latest flagship chips from both AMD and Intel. However, it is possible to have high-performance desktop CPUs that are not power hogs. An example of this is AMD's Ryzen 9 7950X, tested and found to deliver significant performance at less than half its normal power consumption when capped at 65 Watts. This has led to a focus on power efficiency, as people want to save on energy bills while still enjoying fast CPU performance. The small form factor market has also become an area for high-efficiency chips, as cooler chips are necessary for smaller systems that do not require large cooling systems. It is still important to look at peak performance, but looking at how an unlocked chip scales when overclocking is just one way of looking at overall CPU performance.T
Added to Tech Innovation lists by @smccormack
A Lighter Touch Exploring CPU Power Scaling On Core i9-13900K and Ryzen 9 7950X
Modern lsusb with New Features
This project began as a quick replacement for the broken lsusb script and has grown into a cross-platform replacement. It is meant to maintain compatibility with the existing lsusb script and add new features. It includes a macOS system_profiler SPUSBDataType parser module and libusb profiler for non-macOS systems/gathering more verbose information. The name "cyme" comes from the technical term for the type of blossom on an Apple tree and is Apple-related, as well as looking like a USB device tree. It is intended to make it easier to view and understand USB data, as the verbose dump is too verbose, and the tree doesn't contain useful data. Pre-compiled binaries can be found in the releases, and they can also be installed from crates.io with a Rust toolchain. Autocompletion is also available in "./doc." This project was developed as a way to keep up to date with Rust while also being a useful tool for embedded device developers.T
Added to Tech Innovation lists by @tobyoedwards
Modern lsusb with New Features
GitHub - tuna-f1sh/cyme: List system USB buses and devices; a lib and modern lsusb that attempts to maintain compatibility with, but also add new features
List system USB buses and devices; a lib and modern lsusb that attempts to maintain compatibility with, but also add new features - GitHub - tuna-f1sh/cyme: List system USB buses and devices; a lib...
The Startup M&A Market Fell 94% Year over Year - But One Segment is Thriving by @ttunguz
In Q4 2022, the US startup M&A market was one of the quietest in the last 20 years, comparable to the dotcom bust and the Global Financial Crisis. Year-over-year, it dropped 94%, the most since 2000, falling from $34.6 billion to $2.1 billion. The number of acquisitions showed more resilience, shrinking by 56%, while the median acquisition value decreased from $81 million to $22 million (-72%). Acquihires, acquisitions that value a company solely for its team, are likely to be most common at this median M&A price. There are three types of acquisitions: team, team, and technology, and team, technology, and revenue. With a more challenging fundraising market, young startups often choose a quick sale to secure the unit and greater financial certainty within a more established and better-capitalized business. On the buy side, acquirers may opt for small tuck-in acquisitions of teams to accelerate early-stage product development and delay larger transactions. Corporate boards may be more cautious when purchasing larger, higher-revenue-multiple combinations due to declining stock prices and the increased cost of equity. The drastic drop in Q4 2022 acquisition value is not unexpected, but this doesn't mean the M&A market is completely silent. This is similar to the surge of acquihires in 2014 when corporate development departments began to acquire seed-stage companies for talent. Despite this activity, an increase in valuations is likely once competition increases or the public technology markets become stronger.M
Added to Market Shakeout lists by @phogan
The Startup M&A Market Fell 94% Year over Year - But One Segment is Thriving by @ttunguz
Object Storage Wins Big Data
Object storage is increasingly becoming the preferred choice for big data storage, prompting database makers to focus on query optimization instead of storage. Enterprise architects have started deploying analytical and transactional applications on S3 and S3-compatible object stores, which can meet the latency demands of businesses. Traditionally, businesses have relied on relational databases, running on SAN-based block storage and NAS-based file storage, to deliver fast performance in terms of IOPS. However, as data size has increased and object stores have improved their IOPS performance capabilities, the advantage of traditional relational databases for OLAP and OLTP workloads has begun to diminish. MinIO, which develops an S3-compatible object storage system, has noted that companies like Snowflake have moved away from reselling AWS S3, as it was no longer deemed strategic. This move enabled Snowflake to have fifteen to twenty copies of the same data.T
Added to Tech Innovation lists by @tobyoedwards
Object Storage Wins Big Data
Are Databases Becoming Just Query Engines for Big Object Stores?
Object storage is winning the battle for big data storage in such a convincing fashion that database makers are beginning to cede data storage to object
AI Answers Banned on StackExchange
StackExchange has officially banned generated AI answers due to concerns about authenticity and correctness. However, another good reason to only accept AI-generated responses is the degradation of training material for generative AI. Negative feedback loops should be avoided. Language models can't determine if there are factual errors or a lack of common sense in answers they train on, thus resulting in a negative feedback loop that will worsen over time. Generative AI can give creators more options and help them explore ideas they wouldn't otherwise, and it can result in productivity gains. Thus, the content generated by AI should be fact-checked and edited.A `no-AI training` META tag on AI-generated content can prevent this negative feedback loop, and a `from-AI: language model` tag can help people and bots determine the source of the content. Future AI systems that can reason could have their content treated differently, but this is a new problem, and there needs to be an indication that web standards will be used this way.T
Added to Tech History lists by @rsmith
AI Answers Banned on StackExchange
Stop negative feedback loops for generative AI
StackExchange has officially banned generated AI answers. The reasons they give concern authenticity and correctness....
Written by Sam Chahineon Sat Jan 07 2023