<rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hacker News: vamsiraju</title><link>https://news.ycombinator.com/user?id=vamsiraju</link><description>Hacker News RSS</description><docs>https://hnrss.org/</docs><generator>hnrss v2.1.1</generator><lastBuildDate>Sun, 12 Jul 2026 22:42:39 +0000</lastBuildDate><atom:link href="https://hnrss.org/user?id=vamsiraju" rel="self" type="application/rss+xml"></atom:link><item><title><![CDATA[New comment by vamsiraju in "GPT-5.6"]]></title><description><![CDATA[
<p>I think 5.6 Sol is only as good as 5.5 or Opus 4.8 in terms of getting its given work done. It just has an uncanny ability to pickup more work that it can tackle next that the older models lack, or have not been trained to do before.
Where folks are seeing a difference between working with Fable or 5.6 I think also boils down to this phase shift.</p>
]]></description><pubDate>Thu, 09 Jul 2026 19:24:51 +0000</pubDate><link>https://news.ycombinator.com/item?id=48851154</link><dc:creator>vamsiraju</dc:creator><comments>https://news.ycombinator.com/item?id=48851154</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48851154</guid></item><item><title><![CDATA[New comment by vamsiraju in "GPT-5.6"]]></title><description><![CDATA[
<p>I think this is the phase shift 5.6 (Sol set to Ultra) is bringing to the table. Until now we have become accustomed to asking models to continue and their natural inclination is always to stop.
Now OpenAI have flipped it around and for the first time are asking us to steer or stop the model instead, and its own inclination is to keep going. We now have to decide when we need to steer or want to catch up on our understanding of the work done but it will keep going.</p>
]]></description><pubDate>Thu, 09 Jul 2026 19:01:59 +0000</pubDate><link>https://news.ycombinator.com/item?id=48850835</link><dc:creator>vamsiraju</dc:creator><comments>https://news.ycombinator.com/item?id=48850835</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48850835</guid></item><item><title><![CDATA[New comment by vamsiraju in "GPT-5.6"]]></title><description><![CDATA[
<p>prompts -> loops -> slingshots?<p>Its an extremely capable model. I think the way we need to approach works shifts again. We need to get our harnesses/workflows to let it gather some momentum on the first couple rounds but then we also need to structure it so that it can slingshot and accomplish the long range goal.</p>
]]></description><pubDate>Thu, 09 Jul 2026 17:58:33 +0000</pubDate><link>https://news.ycombinator.com/item?id=48849972</link><dc:creator>vamsiraju</dc:creator><comments>https://news.ycombinator.com/item?id=48849972</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=48849972</guid></item><item><title><![CDATA[New comment by vamsiraju in "Notion 3.0"]]></title><description><![CDATA[
<p>Anytype is a great alternative. You can use it in its simple form or design your own workflows with custom document types and properties. In Notion everything is a page but in Anytype you can define document types like Research Note, Concept, Bug Report,.. to distinguish between the various types of pages we typically use to organize information in a space. It's also fully e2e encrypted so you do not have to worry about them using your information to train models. It supports multiple devices and can sync between them p2p while maintaining a hosted server copy but does not have a webapp.</p>
]]></description><pubDate>Tue, 23 Sep 2025 15:26:18 +0000</pubDate><link>https://news.ycombinator.com/item?id=45348402</link><dc:creator>vamsiraju</dc:creator><comments>https://news.ycombinator.com/item?id=45348402</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45348402</guid></item><item><title><![CDATA[New comment by vamsiraju in "Slow Liquid"]]></title><description><![CDATA[
<p>Toggling Reduced Transparency ON in Accessibility makes it so much better.</p>
]]></description><pubDate>Fri, 19 Sep 2025 15:35:13 +0000</pubDate><link>https://news.ycombinator.com/item?id=45302790</link><dc:creator>vamsiraju</dc:creator><comments>https://news.ycombinator.com/item?id=45302790</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=45302790</guid></item><item><title><![CDATA[New comment by vamsiraju in "How middlemen are gaming the H-1B program"]]></title><description><![CDATA[
<p>Multiple registration became an issue for a couple years but the root cause for the large number of applications this year which will continue to happen is because a small change USCIS made in their application process recently which made the application process more democratic and accessible.<p>Prior to 2021, everyone had to submit a full application packet along with the application fees to enter the lottery. Which meant an IT staffing company had to spend ~2000 USD on immigration lawyers to prepare the application for one employee and another cheque of ~2500 USD to USCIS as application fees. The application fees would get refunded back from USCIS if the application was not picked in the lottery for 85,000 slots and the IT Staffing company would have to at minimum bear the costs they spend on lawyers preparing an application they had to submit to enter the lottery.<p>From 2022, USCIS I believe made a change with a good intention to reduce burden on the large number of applicants whose applications were bound to not get picked in the lottery. USCIS always received about 225,000 applications typically every year and the 85,000 slots allocated for the full year would get filled in the first week itself. USCIS made a change to the application process and allowed IT staffing companies to pre-register for the lottery by paying a nominal fee of 10 USD. They now do not have to bear any lawyer costs to enter the lottery since they do not have to submit a full application upfront.<p>This small change makes the whole process more democratic to a vast number of people who work at IT firms in India and allows them to apply now since they fully meet the specialized work conditions set by the H1-B visa system and the cost to enter is almost non-existent.</p>
]]></description><pubDate>Thu, 01 Aug 2024 00:39:26 +0000</pubDate><link>https://news.ycombinator.com/item?id=41125089</link><dc:creator>vamsiraju</dc:creator><comments>https://news.ycombinator.com/item?id=41125089</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=41125089</guid></item><item><title><![CDATA[New comment by vamsiraju in "New Kid on the Blockchain"]]></title><description><![CDATA[
<p>"Maker is a DAO based on the Ethereum blockchain, where it maintains the infrastructure for the Dai - a cryptocurrency that uses smart contract-enforced monetary policy to create price stability.<p>The Dai enables Ethereum users to transact with smart contracts and applications on the Ethereum blockchain without having to deal with the volatility that is associated with traditional cryptocurrencies such as Bitcoin. It also functions as a decentralized savings account as it is long term deflationary, resulting in slow and steady capital gains.<p>Maker is able to maintain the price stability of the Dai through the Dai Credit System, which backs the Dai with collateral stored in Ethereum smart contracts, while simultaneously functioning as an internet-based, p2p credit market that commoditizes credit by allowing anyone with valid collateral to take out loans that have low transaction costs and no middle man fees."<p>[0] <a href="http://makerdao.github.io/docs/" rel="nofollow">http://makerdao.github.io/docs/</a></p>
]]></description><pubDate>Mon, 28 Mar 2016 08:26:03 +0000</pubDate><link>https://news.ycombinator.com/item?id=11373236</link><dc:creator>vamsiraju</dc:creator><comments>https://news.ycombinator.com/item?id=11373236</comments><guid isPermaLink="false">https://news.ycombinator.com/item?id=11373236</guid></item></channel></rss>