TLDR 2025-01-28
Inside DeepSeek π€, iPhone SE 4 leaks π±, Nvidia's $600B loss π
Unlocking the ROI of GRC (Sponsor)
According to
IDC's recent analysis, Vanta delivers an average of $535,000 in annual benefits and a 526% return on investment by automating compliance, streamlining security tasks, and centralizing visibility across GRC programs.
Join the live event on Feb 27 to discover how Vanta empowers organizations to achieve exceptional results in their Governance, Risk, and Compliance (GRC) programs.
Save your spot.
iPhone SE 4 appears in new photos and video, notch and all (2 minute read)
A leaker shared a short video of the new iPhone SE over the weekend. Unlike previous models, the device appears to only have a single rear camera. It is the first iPhone SE model to feature a USB-C port. The device uses Apple's older notched display rather than the Dynamic Island design. Rumors point to a launch around March or April. The video is available in the article.
DeepSeek FAQ (37 minute read)
DeepSeek's R1 model, which is similar to OpenAI's o1, caused a meltdown in the tech industry over the weekend. Many of the revelations that contributed to the meltdown were included in DeepSeek's announcement over Christmas and most of the breakthroughs in V3 were actually revealed with the release of the V2 model last January. DeepSeek-V2 introduced two important breakthroughs, DeepSeekMoE and DeepSeekMLA - the implications of these breakthroughs only became apparent with V3, which added a new approach to load balancing and multi-token prediction in training. The cost of training the model amounted to only $5.576 million, but that excludes the costs associated with prior research, experiments, and data. Many of the decisions DeepSeek made in training its model were due to the constraints of the chip ban - if DeepSeek had access to more powerful chips, it probably would have used a different method.
π
Science & Futuristic Technology
Huge Study Finds Constellation of Health Benefits for Ozempic Beyond Weight Loss (3 minute read)
A new study that tracked millions of diabetes patient outcomes over a period of 3.5 years found that those on GLP-1 medications saw significant improvements across a staggering range of health concerns, far beyond anything clearly linked to weight or blood sugar. The patients experienced improvements in addiction and psychotic disorders and the risk of Alzheimer's was decreased. GLP-1RA drugs act on receptors that are expressed in brain areas involved in impulse control, reward, and addiction, which may potentially explain their effectiveness in curbing appetite and addiction disorders. While the findings are positive, the researchers found some issues that seemed to be exacerbated by taking GLP-1s, including an increase in risk for arthritis and pancreatitis.
With successful New Glenn flight, Blue Origin may finally be turning the corner (13 minute read)
Blue Origin is finally a real rocket company with the successful launch of New Glenn. The company's next steps will be to get better at building engines and rockets and flying New Glenn regularly. Its current priorities are rate manufacturing and driving urgency around building the 'machine that builds the machine'. Blue Origin is the world's largest single-donor nonprofit, so it has zero incentive to operate like SpaceX.
π»
Programming, Design & Data Science
Free Whitepaper: Learn How BOAT Platforms Are Saving Your Competition Thousands of Hours (Sponsor)
Business Orchestration and Automation Technologies (BOAT) combine business process automation, robotic process automation, and integration PaaS into a single solution β making hundreds of manual tasks obsolete.
Camunda's whitepaper reveals how BOAT platforms unify systems/streamline workflows, why the market embraces orchestration platforms, and what the best process architecture looks like.
Get the whitepaperMicrosoft builds open source document database on PostgreSQL, suggests FerretDB as front end (3 minute read)
Developers looking to build something new on a document database now have more alternatives to choose from. Microsoft has launched a document database platform constructed on a relational PostgreSQL back end. It has suggested that users kick things off by implementing a FerretDB solution on the platform to get started. Microsoft has been increasingly accepting of the open source world. The joint development between Microsoft, PostgreSQL, and FerretDB is a threat to source-available MongoDB and Couchbase.
My failed attempt to shrink all npm packages by 5% (12 minute read)
Developer Evan Hahn proposed an idea in 2022 that could decrease the size of all newly-published npm packages by about 5%. Despite it being completely backwards compatible, improving performance, and reducing storage costs, the npm maintainers rejected the proposal. Hahn had thought the idea was great until he received feedback - implementing his idea would be difficult and would make publishing much slower. The changes would also not retroactively apply to existing packages. After receiving feedback and thinking about the tradeoffs, Hahn felt that the idea wasn't worth it and closed the proposal.
Meta is reportedly scrambling βwar rooms' of engineers to figure out how DeepSeek's AI is beating everyone else at a fraction of the price (2 minute read)
Meta's Mark Zuckerberg responded to the release of DeepSeek R1 by assembling four war rooms of engineers to figure out how the Chinese hedge fund built the AI. DeepSeek's model may potentially outperform the next version of Meta's Llama AI due to be released in early 2025. Meta is working on deciphering how DeepSeek lowered the cost of training and running its model, finding out what data the Chinese lab used to train its model, and how Meta can restructure its models based on the attributes of the DeepSeek models. Zuckerberg recently announced that Meta would spend as much as $65 billion on AI-related projects in the coming year, just days after OpenAI, SoftBank, Oracle, and others announced a $500 billion White-house backed AI infrastructure project.
Nvidia sheds almost $600 billion in market cap, biggest one-day loss in U.S. history (4 minute read)
Nvidia experienced the biggest market cap drop for any company on a single day in US history on Monday when it lost close to $600 billion in market cap. Its stock price plummeted 17% and caused a 3.1% slide in the tech-heavy NASDAQ. The sell-off was sparked by DeepSeek presenting increased competition in the global AI industry. The Chinese AI lab released a free and open source large language model that only took two months and less than $6 million to build using Nvidia H800s. The release has caused investors to rethink spending on compute.
Get the most interesting stories in startups, tech, and programming delivered in a free daily email.
Join 1,600,000 readers for
one daily email