DEV Community: AI Tech Connect

RAG in Production 2026: Hybrid Search, Parent-Child, the 40% Bug

AI Tech Connect — Sun, 24 May 2026 15:30:15 +0000

Originally published on AI Tech Connect.

The 40% bug nobody puts on the demo slide Most RAG demos look excellent. You type a question, the system reaches into a vector database, finds a relevant passage and the LLM writes a confident, well-cited answer. Demo over, deal closed, retrospective written. The trouble starts about three weeks later when a real user types a real question and gets a confident, well-cited answer that is built on the wrong document — and nobody on the team realises until a customer complaint lands. The honest number, repeated quietly inside production teams and now written up in the May 2026 production guide from lushbinary, is this: naive RAG pipelines fail at retrieval roughly 40% of the time. The LLM still answers. It is just answering from the wrong evidence. If you are a builder in Bengaluru shipping…

Read the full article on AI Tech Connect →

Open-Weight Coding Shoot-Out: DeepSeek V4 vs Mistral 3.5 vs Gemma 4

AI Tech Connect — Sun, 24 May 2026 13:45:10 +0000

Originally published on AI Tech Connect.

Thirty days, five frontier open-weight models Between late April and the third week of May 2026, five frontier-class open-weight models shipped — one each from DeepSeek, Mistral, Meta, Alibaba and Google. Any one of these releases would have been the story of the month a year ago. Stacked into a single window, they collectively close the open-weight gap to the strongest closed-weight coders (Claude Sonnet 4 / 4.6, GPT-5, Gemini 2.5) and in one case overtake them on the most-cited coding benchmark in the field. For builders in Bengaluru, Pune, London or Manchester evaluating whether to self-host in 2026, the question is no longer "are open-weight models good enough?" It is "which open-weight model fits the shape of my workload, my hardware budget and my licence constraints?" Each of the…

Read the full article on AI Tech Connect →

DPDP Phase II — Six-Month AI Compliance Playbook for India

AI Tech Connect — Sun, 24 May 2026 11:44:11 +0000

Originally published on AI Tech Connect.

Where India actually is on DPDP, today The Digital Personal Data Protection Act and the Digital Personal Data Protection Rules, 2025 were notified by the Ministry of Electronics and Information Technology on 13 November 2025. That single date set in motion a three-phase rollout that ends in May 2027. As of late May 2026, we are squarely in the middle. Here is the timeline, stripped of jargon: Phase I — 13 November 2025 (done). The provisions that stand up the Data Protection Board of India became effective the day the Rules were notified. The Board exists, it is staffed, and it is the body that will adjudicate complaints and impose penalties. Phase II — 13 November 2026 (the six-month window). Consent Manager rules become effective, twelve months from notification. Significant Data…

Read the full article on AI Tech Connect →

DeepSeek V4 Pro Hits 80.6% on SWE-Bench: The New Open-Weight King

AI Tech Connect — Sun, 24 May 2026 10:38:34 +0000

Originally published on AI Tech Connect.

The headline number and why it matters On the SWE-Bench Verified leaderboard, DeepSeek V4 Pro now sits at 80.6% — clear of every other open-weight model and ahead of all the closed-source frontier coders shipping in May 2026. Claude Sonnet 4 sits at roughly 77.2%, GPT-5 at about 74.9%, Gemini 2.5 around 71.8%. On GPQA Diamond, V4 Pro scores 90.1, putting it within striking distance of the top closed-source reasoning models. And it runs on a 1M-token context window. You can download the weights today, point them at your own GPU cluster, and never send a single byte of source code to a third-party API. That last sentence is the whole article, really. The benchmark lead will move within weeks — open-weight models leapfrog each other constantly, and Llama 4, Qwen 3.5, Gemma 4 and Mistral…

Read the full article on AI Tech Connect →

Sierra Raises $950M at $15B: The Customer-Agent Category Arrives

AI Tech Connect — Sun, 24 May 2026 07:30:15 +0000

Originally published on AI Tech Connect.

The round, in one paragraph — then the part that actually matters Sierra, the conversational-agent company founded by former Salesforce co-CEO Bret Taylor, has closed a $950M round led by Tiger Global and GV, pushing the post-money valuation above $15B. TechCrunch reported the deal on 4 May 2026. More than 40% of the Fortune 50 are already paying customers; Sierra's agents handle billions of interactions a year on its platform. For the round mechanics and infrastructure-thesis breakdown, see our earlier coverage of the $950M Series infrastructure round. This piece is a different story. The mechanics are not the lesson. The lesson is that the customer-agent category has now arrived as a standalone software market — not as a feature of a CRM, not as a chatbot tab inside a contact-centre…

Read the full article on AI Tech Connect →

Cerebras IPO at $56B: Why OpenAI Bet 750MW on Wafer-Scale

AI Tech Connect — Sun, 24 May 2026 05:30:13 +0000

Originally published on AI Tech Connect.

The structural shift, not the IPO mechanics Cerebras Systems went public on 14 May 2026, closing its first trading day at a fully diluted valuation of roughly $56 billion. The headline is easy to write and easy to misread. The actual story is not that a niche AI hardware company finally listed; it is what the company committed to in the weeks before pricing — a multi-year deal with OpenAI to deliver up to 750 megawatts of inference capacity over three years on Cerebras's wafer-scale engines. That single anchor customer is what the public-market valuation is really pricing. And that contract is the visible edge of something larger: inference is splitting away from training as a separate market, with separate buyers, separate economics and increasingly separate silicon. Nvidia's GPUs still…

Read the full article on AI Tech Connect →

AISI on Isambard-AI: What UK's Sovereign Supercomputer Is Running

AI Tech Connect — Sat, 23 May 2026 15:30:12 +0000

Originally published on AI Tech Connect.

What you need to know Isambard-AI is now AISI's primary evaluation rig — 21 exaflops of AI performance, 5,400 NVIDIA GH200 Grace Hopper Superchips, sitting in Bristol on a 5-megawatt power envelope. Built with £225 million of UK Government funding and backed by the £500m UK Sovereign AI Fund. The first published case study is on alignment — arXiv 2604.00788 covers how AISI tested whether frontier models would sabotage safety research when placed inside a simulated frontier AI lab. 270 scenarios, four subject models — Claude Opus 4.1, Claude Sonnet 4.5, GPT-5 and a pre-release snapshot of Claude Opus 4.5. The evaluator was Sonnet 4.6, simulating tool responses against eight real research codebases rewritten to look like internal Anthropic projects. The threat model is "sabotage from the…

Read the full article on AI Tech Connect →

Krutrim's Bodhi-1: Can India's First AI Chip Dent NVIDIA Dependence?

AI Tech Connect — Sat, 23 May 2026 13:30:12 +0000

Originally published on AI Tech Connect.

Where Krutrim's silicon stack stands going into 2026 The Krutrim chip announcement is not new — Bhavish Aggarwal unveiled the Bodhi, Sarv and Ojas lines back in 2024 at the Sankalp event at Ola's Futurefactory in Krishnagiri, Tamil Nadu. What is new is the proximity of the 2026 launch window. With six to seven months left on the announced timetable, the conversation has shifted from "can India design an AI chip?" to a much harder question: even if Bodhi-1 ships, can the economics actually compete with NVIDIA imports, and does it move the needle on India's GPU import dependence? Krutrim, Ola Group's AI subsidiary, became India's first AI unicorn after its earlier billion-dollar valuation round. It already runs models, inference services and a small cloud footprint. The chip is the part of…

Read the full article on AI Tech Connect →

B200 vs H100 Inference Economics: When Self-Hosting Wins in 2026

AI Tech Connect — Sat, 23 May 2026 11:30:10 +0000

Originally published on AI Tech Connect.

TL;DR: when does self-hosting beat the API in 2026? Three quick rules, then the tables. Self-host wins when your inference workload is sustained, predictable and large enough to keep at least one B200 above ~50% utilisation 24/7. Cost-per-million-tokens collapses to around $0.02 at that point. API wins when traffic is bursty, when you swap models more than once a quarter, or when you cannot dedicate a platform engineer to the serving stack. Hybrid wins for everyone else. Run the steady-state slice on a reserved B200; spill the long tail to an API or to TPU spot capacity. Pro tip Run the utilisation calculation before the price calculation. A B200 reserved at $2.25/hr ($1,650/month, 24/7) only beats a per-token API once you can keep it meaningfully busy. Idle silicon is the most expensive…

Read the full article on AI Tech Connect →

Microsoft Agent Framework 1.0: AutoGen and Semantic Kernel Merge

AI Tech Connect — Sat, 23 May 2026 09:30:12 +0000

Originally published on AI Tech Connect.

What changed on 3 April One SDK, not two. The enterprise foundations of Semantic Kernel and the multi-agent orchestration patterns of AutoGen now ship as a single open-source package under the Microsoft.Agents.AI namespace. Same concepts across .NET and Python. The API shape is mirrored across both languages — a rare commitment from Microsoft to keep its polyglot story straight. Six model providers out of the box. Azure OpenAI, OpenAI, Anthropic Claude, Amazon Bedrock, Google Gemini and Ollama. The framework is no longer Azure-only in any meaningful sense. MCP and A2A are first-class. Model Context Protocol for tool discovery, Agent-to-Agent for cross-runtime collaboration — both baked in rather than bolted on. DevUI and OpenTelemetry ship by default. Browser-based local debugger plus…

Read the full article on AI Tech Connect →

DeepSeek's First VC Round: What a $45B Tag Says About China-AI Capital

AI Tech Connect — Sat, 23 May 2026 07:30:09 +0000

Originally published on AI Tech Connect.

The deal in one line DeepSeek — the Hangzhou lab that has spent the last eighteen months handing the open-weight community a free seat at the frontier — is reportedly in talks to raise between $3 billion and $4 billion at a valuation of roughly $45 billion, with some outlets putting the upper end nearer $50 billion. The lead investor in discussions is the China Integrated Circuit Industry Investment Fund, the state-backed vehicle the country's tech press still simply calls "Big Fund". Tencent and Hillhouse Capital are named as likely co-investors. TechCrunch broke the story on 6 May; Bloomberg and TechNode have since corroborated the broad shape of the round. For a company that has, up until now, financed every training run out of the pocket of its hedge-fund parent High-Flyer, this is a…

Read the full article on AI Tech Connect →

Cursor Composer 2.5 Hits Opus 4.7 and GPT-5.5 Parity at Half the Price

AI Tech Connect — Sat, 23 May 2026 05:30:09 +0000

Originally published on AI Tech Connect.

What this means for coding-agent buyers On 18 May 2026, Cursor shipped Composer 2.5 — the second iteration of its in-house coding model — and the colour of the conversation around coding agents changed overnight. The headline is not just another benchmark win. It is the price tag. Frontier-class quality at a tenth of the price — Composer 2.5 sits inside a percentage point of Opus 4.7 on SWE-Bench Multilingual and edges out GPT-5.5 on CursorBench v3.1 default-effort. Standard-tier input at $0.50 per million tokens — that is 10× cheaper per token than Opus 4.7 list pricing. For high-volume batch and background-agent work, the unit economics no longer look like a frontier model. Independent indices placed Cursor near the top of the coding-agent table at launch — with a cost-per-task curve…

Read the full article on AI Tech Connect →