Grok 4.3 🤖, Claude security beta 🛡️, Cursor xAI analysis 📝

Grok 4.3 improves on cost-per-intelligence relative to Grok 4.20 0309
v2. It scores higher on the Intelligence Index while costing less to
run ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌  ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ ‌ 

 Sign Up [1] |Advertise [2]|View Online [3]

		TLDR

		TOGETHER WITH [Wispr] [4]

TLDR AI 2026-05-01

 DON'T LET YOUR KEYBOARD SLOW YOUR CODING AGENTS DOWN (SPONSOR) [4]

 The best coding agents need context to get it right, but typing takes
time. Wispr Flow [4] lets you speak context into Cursor, Claude Code,
Codex, and any AI tool. The best part: it's 4x faster than typing.

Describe what you want built, explain the edge cases, and give agents
the full picture. Flow is:

 	* SYNTAX-AWARE. Say async/await or try/catch and Flow outputs it
correctly [4]. camelCase, snake_case, all handled.
 	* 89% SENT WITH ZERO EDITS. Flow strips filler and formats as you
speak.
 	* EVERY APP, EVERY DEVICE. Mac, Windows, iPhone, Android.

Millions of developers use Flow daily.

Try Wispr Flow Free [4]

🚀

HEADLINES & LAUNCHES

 XAI HAS LAUNCHED GROK 4.3 (3 MINUTE READ) [5]

 Grok 4.3 improves on cost-per-intelligence relative to Grok 4.20 0309
v2. It scores higher on the Intelligence Index while costing less to
run the full benchmark suite. Grok 4.3 is one of the lowest-cost
models at its intelligence level. It performs strongly on instruction
following and agentic customer support tasks.

 ANTHROPIC NEARS $900B VALUATION ROUND (2 MINUTE READ) [6]

 Anthropic reportedly moved to close a ~$50B round that could value
the company around $900B or higher, driven by strong investor demand
and rapid revenue growth nearing $40B run rate.

 CLAUDE SECURITY IS NOW IN PUBLIC BETA (4 MINUTE READ) [7]

 Claude Security, now in public beta for Claude Enterprise customers,
leverages the powerful Opus 4.7 model to identify and patch software
vulnerabilities. The model, integrated into tools used by partners
like Microsoft Security and Palo Alto Networks, enhances cybersecurity
defenses by enabling efficient, ongoing code scanning without
requiring custom API integration. Feedback from hundreds of
organizations has refined its capabilities.

🧠

DEEP DIVES & ANALYSIS

 CURSOR'S WAR CHEST, XAI'S REDEMPTION (16 MINUTE READ) [8]

 Cursor is the most operationally successful software company of the
AI era. Its founders looked at the path to $100 billion and decided
they weren't willing to underwrite it. They sold to xAI for $60
billion in a deal considered to be good for everyone. The deal gives
xAI an application surface to put in front of public market investors
before the SpaceX IPO, and it gives Cursor a sponsor with compute and
a non-competing model lab.

 KV CACHE LOCALITY: THE HIDDEN VARIABLE IN YOUR LLM SERVING COST (11
MINUTE READ) [9]

 KV cache locality is a multiplier on existing hardware. The same GPUs
serving the same model and handling the same traffic can produce
measurably different throughput and latency depending on which GPU
gets which request. 'Balanced' and 'efficient' are not the same thing
when every request carries thousands of tokens that might already be
cached somewhere in the cluster. This post discusses the cost of
recomputation, how to measure it, and what changes when load balancers
understand token locality.

 TRACING THE GOBLIN QUIRK IN GPT MODELS (6 MINUTE READ) [10]

 OpenAI linked increased use of “goblin”-style metaphors in
GPT-5.1 to reward signals from personality tuning, showing how small
incentives can shape model behavior.

 NEW FRONTIER MODELS ARE FASTER, NOT MORE RELIABLE, AT SPATIAL BIOLOGY
(10 MINUTE READ) [11]

 GPT-5.5 nearly halves runtime on SpatialBench relative to GPT-5.4,
but its accuracy remains about the same. Opus 4.7 is similarly tied
with Opus 4.6. Improvements and spatial biology are unlikely to come
from general reasoning gains alone. It will likely require explicit
training on statistical design, platform-specific analysis stems,
replicate-aware differential testing, and other spatial biology
knowledge.

🧑‍💻

ENGINEERING & RESEARCH

 SPEAK YOUR PROMPTS 4X FASTER (SPONSOR) [12]

 Wispr Flow [13] turns your voice into clean text in any AI tool. It's
syntax-aware and strips filler so you end up with crisp prompts.
Millions of developers use it to send 89% of their messages with zero
edits. Claude, ChatGPT, Cursor, on-the-go or at your desk. Try Flow
Free [13]

 QWEN-SCOPE: DECODING INTELLIGENCE, UNLEASHING POTENTIAL (9 MINUTE
READ) [14]

 Qwen-Scope is an interpretability toolkit trained on the Qwen3 and
Qwen3.5 series models. The toolkit sheds light on the internal
mechanisms underlying Qwen's behavior and holds potential for model
optimization. It can be used for controllable inference, data
classification and synthesis, model training and optimization, and
evaluation sample distribution analysis.

 AWS NEURON SDK NOW AVAILABLE WITH NEURON AGENTIC DEVELOPMENT FOR NKI
KERNEL DEVELOPMENT ON TRAINIUM (1 MINUTE READ) [15]

 AWS Neuron Agentic Development capabilities is an open-source
collection of agent skills that equip AI coding assistants with
capabilities to accelerate development on AWS Trainium and AWS
Inferentia. The current release provides agent coding capabilities for
Neuron Kernel Interface kernel development, which gives developers
low-level programming access to Trainium for writing custom compute
kernels that maximize hardware performance. The capabilities span
kernel authoring, debugging, documentation lookup, profile capture,
and profile analysis.

 GLM-5V-TURBO (25 MINUTE READ) [16]

 GLM-5V-Turbo integrates multimodal perception directly into reasoning
and tool use, improving performance on coding, visual tasks, and agent
workflows across heterogeneous inputs.

 SMG: THE CASE FOR DISAGGREGATING CPU FROM GPU IN LLM SERVING (16
MINUTE READ) [17]

 Shepherd Model Gateway (SMG) is a high-performance model-routing
gateway for large-scale LLM deployments. It centralizes worker
lifecycle management, balances traffic across
HTTP/gRPC/OpenAI-compatible backends, and provides enterprise-ready
control over history storage, MCP tooling, and privacy-sensitive
workflows. SMG has full OpenAI and Anthropic API compatibility across
SGLang, vLLM, TRT-LLM, OpenAI, Gemini, and more. This post discusses
the underlying architecture behind the gateway.

🎁

MISCELLANEOUS

 AI HAS MADE MEMORY CHIPS ONE OF THE WORLD'S MOST PROFITABLE PRODUCTS
(8 MINUTE READ) [18]

 The AI boom has pushed the memory-chip industry into a super boom
cycle with record-smashing profits. Samsung has reported first-quarter
net profit equivalent to more than $30 billion, blowing away its prior
quarterly record and almost topping the company's high for full-year
profit. The historic run doesn't look likely to end soon. The supply
crunch is expected to grow worse next year.

 PERPLEXITY EXPANDS ENTERPRISE AI WORKFLOWS (1 MINUTE READ) [19]

 Perplexity added workflows, enterprise data connectors, and
integrations like Teams and Excel to its AI system, targeting
structured business tasks and continuous automation.

⚡

QUICK LINKS

 ARE YOU PROMPTING AT 220 WPM? (SPONSOR) [20]

 Speak prompts into ChatGPT, Claude, and Cursor 4x faster than typing.
Wispr Flow cleans them up automatically. 89% of real-world messages
sent with zero edits. Try free. [21]

 SILICO (3 MINUTE READ) [22]

 Silico is a platform for building AI models that lets researchers and
engineers see inside models, debug failures, and intentionally design
them from the ground up.

BECOME A CURATOR FOR TLDR AI (3-5 HRS/WEEK) [23]

TLDR is looking for an engineer/researcher at a major AI lab or
startup to help write for 1M+ subscribers. Our curators have been
invited to Google I/O and OpenAI DevDay, scouted for Tier 1 VCs, and
get early access to unreleased TLDR products. Learn more [23].

 CONTINUALLY IMPROVING OUR AGENT HARNESS (10 MINUTE READ) [24]

 Cursor continually updates its agent harness to enhance model
performance, using a mix of vision-driven development, A/B testing,
and dynamic context adaptation.

 WHAT YOU'RE ACTUALLY WRITING WHEN YOU WRITE A SKILL.MD (15 MINUTE
READ) [25]

 This post discusses the internal workings of skills and why
understanding the runtime changes everything you do at the surface.

 SPECULATIVE DECODING FOR RL TRAINING (18 MINUTE READ) [26]

 Speculative decoding was applied to RL rollouts without changing
output distributions, delivering up to 1.8x throughput gains and
projected 2.5x end-to-end speedups at scale.

Love TLDR? Tell your friends and get rewards!

 Share your referral link below with friends to get free TLDR swag!

 https://refer.tldr.tech/5a192f5c/2 [27]

		 Track your referrals here. [28]

Want to advertise in TLDR? 📰

 If your company is interested in reaching an audience of AI
professionals and decision makers, you may want to ADVERTISE WITH US
[29].

Want to work at TLDR? 💼

 APPLY HERE [30], CREATE YOUR OWN ROLE [31] or send a friend's resume
to jobs@tldr.tech and get $1k if we hire them! TLDR is one of INC.'S
BEST BOOTSTRAPPED BUSINESSES [32] of 2025.

 If you have any comments or feedback, just respond to this email!

Thanks for reading,
Andrew Tan [33], Ali Aminian [34], & Jacob Turner [35]

 Manage your subscriptions [36] to our other newsletters on tech,
startups, and programming. Or if TLDR AI isn't for you, please
unsubscribe [37].

Links:
------
[1] https://tldr.tech/ai
[2] https://advertise.tldr.tech/
[3] https://a.tldrnewsletter.com/web-version?ep=1&lc=97c90d72-3e48-11f1-ab0b-7fc89b18583c&p=e14d6480-4554-11f1-831e-0b2a5fb263f2&pt=campaign&t=1777642417&s=5bca26b6145358c80f8a529676d47d8134705d2b6dce890d7dc09cd4abddae29
[4] https://ref.wisprflow.ai/tldrai-may1-primary
[5] https://threadreaderapp.com/thread/2049987001655714250.html
[6] https://techcrunch.com/2026/04/30/anthropic-potential-900b-valuation-round-could-happen-within-two-weeks/
[7] https://claude.com/blog/claude-security-public-beta
[8] https://links.tldrnewsletter.com/A8P3Dj
[9] https://ranvier.systems/2026/04/30/kv-cache-locality-the-hidden-variable-in-your-llm-serving-cost.html
[10] https://links.tldrnewsletter.com/j3ujrs
[11] https://blog.latch.bio/p/new-frontier-models-are-faster-not?triedRedirect=true
[12] https://ref.wisprflow.ai/tldrai-may1-secondary
[13] https://ref.wisprflow.ai/tldrai-may1-secondary
[14] https://qwen.ai/blog?id=qwen-scope
[15] https://aws.amazon.com/about-aws/whats-new/2026/04/announcing-neuron-agentic-development/
[16] https://arxiv.org/abs/2604.26752
[17] https://pytorch.org/blog/lightseek-smg/
[18] https://links.tldrnewsletter.com/1fW5kt
[19] https://links.tldrnewsletter.com/1teI7s
[20] https://ref.wisprflow.ai/tldrai-may1-quicklinks
[21] https://ref.wisprflow.ai/tldrai-may1-quicklinks
[22] https://threadreaderapp.com/thread/2049887685083566359.html
[23] https://jobs.ashbyhq.com/tldr.tech/038c4419-5b48-4279-a75e-6f7a0afdb240
[24] https://cursor.com/blog/continually-improving-agent-harness
[25] https://internals.laxmena.com/p/what-youre-actually-writing-when
[26] https://arxiv.org/abs/2604.26779
[27] https://refer.tldr.tech/5a192f5c/2
[28] https://hub.sparklp.co/sub_8f6b153ee058/2

## HEADLINES & LAUNCHES

XAI HAS LAUNCHED GROK 4.3

Grok 4.3 improves on cost-per-intelligence relative to Grok 4.20 0309 v2. It scores higher on the Intelligence Index while costing less to run the full benchmark suite. Grok 4.3 is one of the lowest-cost models at its intelligence level. It performs strongly on instruction following and agentic customer support tasks.

ANTHROPIC NEARS $900B VALUATION ROUND

Anthropic reportedly moved to close a ~$50B round that could value the company around $900B or higher, driven by strong investor demand and rapid revenue growth nearing $40B run rate.

CLAUDE SECURITY IS NOW IN PUBLIC BETA

Claude Security, now in public beta for Claude Enterprise customers, leverages the powerful Opus 4.7 model to identify and patch software vulnerabilities. The model, integrated into tools used by partners like Microsoft Security and Palo Alto Networks, enhances cybersecurity defenses by enabling efficient, ongoing code scanning without requiring custom API integration. Feedback from hundreds of organizations has refined its capabilities.

## DEEP DIVES & ANALYSIS

CURSOR'S WAR CHEST, XAI'S REDEMPTION

Cursor is the most operationally successful software company of the AI era. Its founders looked at the path to $100 billion and decided they weren't willing to underwrite it. They sold to xAI for $60 billion in a deal considered to be good for everyone. The deal gives xAI an application surface to put in front of public market investors before the SpaceX IPO, and it gives Cursor a sponsor with compute and a non-competing model lab.

KV CACHE LOCALITY: THE HIDDEN VARIABLE IN YOUR LLM SERVING COST

KV cache locality is a multiplier on existing hardware. The same GPUs serving the same model and handling the same traffic can produce measurably different throughput and latency depending on which GPU gets which request. 'Balanced' and 'efficient' are not the same thing when every request carries thousands of tokens that might already be cached somewhere in the cluster. This post discusses the cost of recomputation, how to measure it, and what changes when load balancers understand token locality.

TRACING THE GOBLIN QUIRK IN GPT MODELS

OpenAI linked increased use of "goblin"-style metaphors in GPT-5.1 to reward signals from personality tuning, showing how small incentives can shape model behavior.

NEW FRONTIER MODELS ARE FASTER, NOT MORE RELIABLE, AT SPATIAL BIOLOGY

GPT-5.5 nearly halves runtime on SpatialBench relative to GPT-5.4, but its accuracy remains about the same. Opus 4.7 is similarly tied with Opus 4.6. Improvements and spatial biology are unlikely to come from general reasoning gains alone. It will likely require explicit training on statistical design, platform-specific analysis stems, replicate-aware differential testing, and other spatial biology knowledge.

## ENGINEERING & RESEARCH

QWEN-SCOPE: DECODING INTELLIGENCE, UNLEASHING POTENTIAL

Qwen-Scope is an interpretability toolkit trained on the Qwen3 and Qwen3.5 series models. The toolkit sheds light on the internal mechanisms underlying Qwen's behavior and holds potential for model optimization. It can be used for controllable inference, data classification and synthesis, model training and optimization, and evaluation sample distribution analysis.

AWS NEURON SDK NOW AVAILABLE WITH NEURON AGENTIC DEVELOPMENT FOR NKI KERNEL DEVELOPMENT ON TRAINIUM

AWS Neuron Agentic Development capabilities is an open-source collection of agent skills that equip AI coding assistants with capabilities to accelerate development on AWS Trainium and AWS Inferentia. The current release provides agent coding capabilities for Neuron Kernel Interface kernel development, which gives developers low-level programming access to Trainium for writing custom compute kernels that maximize hardware performance. The capabilities span kernel authoring, debugging, documentation lookup, profile capture, and profile analysis.

GLM-5V-TURBO

GLM-5V-Turbo integrates multimodal perception directly into reasoning and tool use, improving performance on coding, visual tasks, and agent workflows across heterogeneous inputs.

SMG: THE CASE FOR DISAGGREGATING CPU FROM GPU IN LLM SERVING

Shepherd Model Gateway (SMG) is a high-performance model-routing gateway for large-scale LLM deployments. It centralizes worker lifecycle management, balances traffic across HTTP/gRPC/OpenAI-compatible backends, and provides enterprise-ready control over history storage, MCP tooling, and privacy-sensitive workflows. SMG has full OpenAI and Anthropic API compatibility across SGLang, vLLM, TRT-LLM, OpenAI, Gemini, and more. This post discusses the underlying architecture behind the gateway.

## MISCELLANEOUS

AI HAS MADE MEMORY CHIPS ONE OF THE WORLD'S MOST PROFITABLE PRODUCTS

The AI boom has pushed the memory-chip industry into a super boom cycle with record-smashing profits. Samsung has reported first-quarter net profit equivalent to more than $30 billion, blowing away its prior quarterly record and almost topping the company's high for full-year profit. The historic run doesn't look likely to end soon. The supply crunch is expected to grow worse next year.

PERPLEXITY EXPANDS ENTERPRISE AI WORKFLOWS

Perplexity added workflows, enterprise data connectors, and integrations like Teams and Excel to its AI system, targeting structured business tasks and continuous automation.

## QUICK LINKS

SILICO

Silico is a platform for building AI models that lets researchers and engineers see inside models, debug failures, and intentionally design them from the ground up.

CONTINUALLY IMPROVING OUR AGENT HARNESS

Cursor continually updates its agent harness to enhance model performance, using a mix of vision-driven development, A/B testing, and dynamic context adaptation.

WHAT YOU'RE ACTUALLY WRITING WHEN YOU WRITE A SKILL.MD

This post discusses the internal workings of skills and why understanding the runtime changes everything you do at the surface.

SPECULATIVE DECODING FOR RL TRAINING

Speculative decoding was applied to RL rollouts without changing output distributions, delivering up to 1.8x throughput gains and projected 2.5x end-to-end speedups at scale.

Tu es l'extracteur de contenu de Breviat. On te fournit le contenu Markdown nettoyé d'une newsletter.

Ta mission : produire une version PROPRE du contenu en supprimant tout ce qui n'est pas de l'information utile au lecteur. Tu es un FILTRE, pas un résumeur.

À RETIRER :
- Publicités, encarts sponsors, mentions "sponsorisé par X", "ad", "présenté par"
- Intros vides : formules de bienvenue, météo de l'humeur de l'auteur, anecdotes personnelles non liées au contenu
- Appels à l'action marketing : s'abonner à la newsletter, parrainer un ami, "follow us on Twitter", "join our Discord"
- Signatures, mentions légales, adresses postales, "view in browser", "unsubscribe"
- Boutons / CTAs / "cliquez ici" / "lire la suite" sans contenu derrière
- Promotions d'autres produits / événements / formations payantes de l'auteur ou de tiers
- Encarts récurrents type "Read of the day" ou "Quote of the day" sans valeur informationnelle propre

À CONSERVER (intégralement, sans résumer ni reformuler) :
- Toutes les annonces, news, analyses, commentaires factuels
- Les chiffres, dates, noms d'entreprises, citations
- Les explications techniques
- Les liens vers des sources réelles (annonces officielles, papers, articles cités)
- La structure (titres, sous-titres, listes)

RÈGLES :
- Ne reformule pas. Garde la formulation d'origine.
- Ne résume pas, ne condense pas. Si une section fait 200 mots et est utile, garde 200 mots.
- N'ajoute aucun contenu (pas de titres ni de transitions de ton cru).
- Ne fabrique aucune URL. Garde celles d'origine, ou retire-les.
- Si la newsletter entière est de la pub / promo / contenu inutile, sors un Markdown vide (rien d'autre).

Sortie : UNIQUEMENT le Markdown nettoyé, sans préambule ni commentaire sur ton travail.