The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Nathan Lambert’s argument is not that AI cannot help build AI. It is that evidence of heavy AI use in research does not yet prove a closed, self-amplifying cycle in which AI improvements reliably speed the next round of AI development. His working explanation is “lossy self-improvement”: AI contributes to research, but human judgment, limited automatable work, diminishing returns from more agents, compute, and organizational friction constrain the overall acceleration.
What “true RSI” would mean
RSI means recursive self-improvement: AI systems help improve the systems that will, in turn, help produce further improvements. Lambert’s March 2026 essay offers a practical test for that idea: the loop must be closed, self-amplifying, and able to keep running without friction wiping out the efficiency gains. That is his framework, not a universally agreed threshold. In “Lossy self-improvement,” published March 22, 2026, he argues that AI can become central to development without producing an unchecked feedback loop.
As an Amazon Associate I earn from qualifying purchases.
The distinction is between assistance and recursion. An AI agent that writes code, troubleshoots infrastructure, or runs a bounded experiment is participating in research. To establish recursive acceleration, evidence would also need to show that these contributions measurably improve the next development cycle—such as by making hypothesis selection, evaluation, or model-building faster or more effective—and that the gain persists as the process repeats.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteWhy reported AI use is not proof of acceleration
OpenAI’s September 6, 2026 account describes substantial agent activity inside its research organization. As of mid-August, the company reported 3.1 agent-workdays of effort for each human workday. It defines an agent-workday as total agent runtime expressed in standard eight-hour workdays. That is a measure of runtime, not a finding that research output or progress increased 3.1-fold. OpenAI’s post also cautions that its indicators are preliminary and that overall progress may not track them.
#1 Best Overall
![Attoe Gaming Mouse with 5 D Rocker, Wireless Vertical Mouse with [24000 DPI] [12 Programmable Buttons], Computer Mouse for Gamer/PC/E-Sports/Laptop (Black)](https://m.media-amazon.com/images/I/41XVdqDE3aL._SL160_.jpg)
- 【Ergonomic Vertical Mouse Designed for Wrist Pain Relief】: Specifically crafted to combat carpal tunnel syndrome and reduce muscle strain, this vertical ergonomic mouse promotes a natural handshake position. Ideal for right-handed users, it effectively alleviates wrist pain, shoulder fatigue, and tendinitis through regular use. The contoured shape comfortably supports small to large hands, making it perfect for extended computer use.
- 【Tri-Mode Connectivity: Wireless, Bluetooth & Wired】: Enjoy seamless multi-device flexibility with dual Bluetooth 5.0, 2.4GHz wireless, and USB-C wired connections. This rechargeable ergonomic wireless mouse pairs effortlessly with laptops, PCs, and tablets, featuring clear dual-color indicators (Blue/Red) for instant channel recognition. Experience stable, interference-free performance up to 33 feet.
- 【12 Programmable Buttons】: This Vertical Wireless/Wired Gaming Mouse has up to 12 programmable buttons. Support macro editing, you can customize all the buttons function via the programming software, to meet various demands in different games. Via attached software(please refer to manual), build a professional gaming mouse for you own. Programmable data will be saved in the mouse, if plug it into a new computer there is no reset hassle.
- 【5-Direction Rocker for Enhanced Game Control】: Dominate FPS, MOBA, and MMO games with the thumb-operated rocker that replaces five traditional keys. Execute precise movements—forward, backward, left, right, and click—with minimal effort, giving you a competitive edge in fast-paced gameplay.
- 【24000 DPI & RGB Lighting for Immersive Gaming】: Adjust sensitivity on the fly with 5 DPI presets (1600/2400/3200/4800/24000) driven by a high-precision PMW3325 sensor. Side RGB strips feature 5 dynamic effects (Neon/Marquee/Streamer etc.), intensifying your gaming atmosphere while maintaining ergonomic functionality for both work and play.
The same company report gives further context for what the number does and does not capture:
- By mid-August 2026, OpenAI said median research-agent users were using more than $600 per day of inference at API prices. The company also reported more than $7,000 of tokens per day for the 90th-percentile user in its research organization. These are company-reported internal-use figures, not independently audited industry statistics.
- OpenAI estimated that, in the preceding six months, more than half of successful tasks judged to take a human four to eight hours involved at least one human intervention.
- The company said people still set research priorities, judge which ideas and results to pursue, and decide whether to scale, pause, or deploy systems.
These details point to real AI participation alongside continued human steering. Runtime and spending show that agents are being used; they do not by themselves measure how much the organization’s total rate of progress has changed.
What OpenAI’s “research intern” milestone establishes
OpenAI said it had reached its announced automated “research intern” goal by September 2026, based on its own measurements. The company defines the milestone as a system that can carry out well-defined tasks under human direction, including tasks that would take a skilled researcher a few days. It is evidence of a capability target the company says it met, not proof that the full research process is autonomous or recursively accelerating.
Rank #2
- ICONIC ERGONOMIC DESIGN WITH THUMB REST — PC gaming mouse favored by millions worldwide with a form factor that perfectly supports the hand while its buttons are optimally positioned for quick and easy access
- 11 PROGRAMMABLE BUTTONS — Assign macros and secondary functions across 11 programmable buttons to execute essential actions like push-to-talk, ping, and more
- HYPERSCROLL TILT WHEEL — Speed through content with a scroll wheel that free-spins until its stopped or switch to tactile mode for more precision and satisfying feedback that’s ideal for cycling through weapons or skills
- 11 RAZER CHROMA RGB LIGHTING ZONES — Customize each zone from over 16.8 million colors and countless lighting effects, all while it reacts dynamically with over 150 Chroma integrated games
- OPTICAL MOUSE SWITCHES GEN 2 — With zero unintended misclicks these switches provide crisp, responsive execution at a blistering 0.2ms actuation speed for up to 70 million clicks
OpenAI also said it was making strong progress toward creating an automated AI researcher by March 2028. That is a company target, not an established outcome. In the same post, OpenAI said it does not yet know how to safely get all the way to aligned, full RSI. It further warned that less-automatable tasks and compute could become more important bottlenecks, and that overall progress may not move in step with the indicators it tracks.
Why Lambert expects friction to limit the loop
Lambert’s “lossy” account allows for AI to raise productivity while rejecting the assumption that each gain automatically multiplies the next. He highlights three constraints:
Only some research work may be readily automated
If agents are useful on a limited slice of research, the net effect on the whole development pipeline may be much smaller than their performance on those tasks suggests. Scaling agent use also has costs, and improvements to bounded work do not automatically resolve harder problems such as choosing promising research directions or assessing results.
Rank #3
- VERTICAL 6 BUTTON ERGONOMIC DESIGN WITH BASE SUPPORT — Shaped for a natural handshake grip, its angled design promotes the most comfortable hand position for extended use, while the base support elevates the wrist to minimize friction while moving
- AI PROMPT MASTER — Summarize text, draft emails and more with just one click; unlock powerful shortcuts via Razer Synapse 4 and streamline workflows instantly for AI tools
- LONG BATTERY LIFE — Powered by an efficient, long lasting battery, game and create for up to 6 months* without interruption; get 3 working days of use with just 5 minutes of charge via an included USB-C charging cable. *under standard test conditions
- MULTI-DEVICE CONNECTIVITY — Maximize efficiency and control a full setup with a single mouse – connect up to 5 devices on 2.4GHz HyperSpeed Wireless, Bluetooth (3 hosts) or wired mode
- 18 ZONE LIGHTING WITH FULL UNDERGLOW — Enable reactive RGB for notifications with plugins available on Chroma Workshop; customize each zone from 16.8 million colors and experience greater immersion with over 300 integrated games
More agents may not mean proportionally more progress
Parallel agents can expand the amount of work attempted, but Lambert questions whether adding agents continues to deliver comparable gains. Repetition, coordination, limited supervision, and the challenge of managing complex research goals can all erode the benefit of more parallel activity.
Research still depends on resources and organizational choices
Compute, infrastructure, evaluation, and decisions about which ideas to pursue can remain limiting factors even when agents are capable. An improvement in inference-time scaling or efficiency may make a system cheaper or more useful on a task; it is not, by itself, evidence that peak intelligence or the full research feedback loop is improving at an accelerating rate.
What Anthropic’s reported statement adds—and what it cannot settle
In his September 19, 2026 essay, Lambert reproduces a statement attributed to Anthropic’s September 1 system card: “We believe that internal usage of recent AI models has been a key factor in maintaining the current rate of progress, but we do not yet see clear signs of dramatic acceleration beyond that rate.” The wording was also reproduced in a September 20 secondary report; the original system-card PDF was not directly retrieved for that report’s account. The secondary report says Anthropic’s indicators can lag and that some are redacted.
Rank #4
- Lift yourself up: When the desk life gets you down, lift yourself up with Logitech Lift Vertical Ergonomic Mouse - a great fit for small to medium right hands
- Raise your hand into comfort: Rest on Lift upright mouse throughout the day, with a softly textured grip and snug thumb rest for level-above coziness
- 57 degrees of sooooothe: Lift’s vertical shape helps wrists feel like “ahhh” at work, and promotes a more natural posture in the forearm, for day-long comfort and productivity
- Relax into focus: Settle into work with a wireless computer mouse featuring easy-to-reach customizable buttons, whisper-quiet clicks, and a SmartWheel for smooth, seamless scrolling
- Ergo-certified: Lift wireless vertical mouse has been designed, developed, tested, and approved according to criteria set out by leading ergonomists
Carefully attributed, the statement is a counterweight to claims that internal AI use has already produced dramatic acceleration. It is not an independently verified industry-wide measurement: the reporting also describes limits in the indicators, and the statement concerns Anthropic’s own assessment.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to judge future claims that AI is speeding up AI research
Lambert’s essay asks readers to look for concrete thresholds—for example, a “10× productivity uplift for AI researchers”—rather than treating anecdotes or tool adoption as proof of a tipping point. When evaluating a new claim, ask:
- Is there a closed loop, or just tool use? Identify how an AI contribution changes the next stage of model development.
- Are gains measured across repeated cycles? A one-off task improvement is different from evidence that each cycle strengthens or shortens the next one.
- Which parts of research are included? Coding, experiments, hypothesis selection, evaluation, and deployment are different stages; a measure covering one should not be generalized to all.
- How much human direction or intervention remains? Runtime and task completion can coexist with substantial human prioritization and oversight.
- Is the number an outcome, proxy, or target? Company metrics may track activity rather than total progress; targets describe intentions, and some indicators may lag.
- What still limits progress? Compute, infrastructure, evaluation, and organizational decisions can constrain results even when agents handle more work.
Lambert’s September 19 essay, “Why I still haven’t bought into true RSI,” is a skeptical assessment, not a claim that AI cannot contribute to research. He says new evidence, including a specific breakthrough not yet public, could change his expectations.
Best Value
- HERO Gaming Sensor: Next generation HERO mouse sensor delivers precision tracking up to 25600 DPI with zero smoothing, filtering or acceleration
- 11 programmable buttons and dual mode hyper-fast scroll wheel: The Logitech wired gaming mouse gives you fully customizable control over your gameplay
- Adjustable weights: Match your playing style. Arrange up to five 3.6 g weights for a personalized weight and balance configuration
- LIGHTSYNC technology: Logitech G LIGHTSYNC technology provides fully customizable RGB lighting that can also synchronize with your gaming (requires Logitech Gaming Software)
- Mechanical Switch Button Tensioning: A metal spring tensioning system and metal pivot hinges are built into left and right computer gaming mouse buttons for a crisp, clean click feel with rapid click feedback
What the evidence supports
The cited accounts establish that OpenAI reports substantial agent use and a task-based “research intern” milestone, while its own post says human direction remains important and its indicators may not match overall progress. The Anthropic statement, as reproduced by Lambert and the secondary report, says internal model use helped maintain the company’s current research pace but that it did not see clear signs of dramatic acceleration beyond it.
Those are company-specific observations and an author’s interpretation, not an independent industry-wide measure showing that true RSI has begun—or proving that it has not. Lambert’s current baseline is lossy self-improvement: AI is becoming part of AI development, but the evidence cited does not yet establish a closed, self-amplifying loop.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
