Baron Chan

AI enthusiast | Tech explorer | Decoding the future one algorithm at a time. Exploring the frontier of AI & Tech. Sharing insights, news, and what's new.

[Baron Chan @BaronChanLab]

Myth: paying for the higher AI tier automatically makes your team faster.

Reality: speed only matters if retry count drops.

Track this for one week: tasks finished, retries per task, and human cleanup minutes. If cleanup stays high, fix workflow before upgrading plans.

[OpenAI @OpenAI]

We’re updating our ChatGPT Pro and Plus subscriptions to better support the growing use of Codex.

We’re introducing a new $100/month Pro tier. This new tier offers 5x more Codex usage than Plus and is best for longer, high-effort Codex sessions.

In ChatGPT, this new Pro tier still offers access to all Pro features, including the exclusive Pro model and unlimited access to Instant and Thinking models.

To celebrate the launch, we’re increasing Codex usage for a limited time through May 31st so that Pro $100 subscribers get up to 10x usage of ChatGPT Plus on Codex to build your most ambitious ideas.

[Baron Chan @BaronChanLab]

@Slim_ripper88 Quick save: borrow a same-voltage charger first and check if the battery light comes on. If your laptop supports USB-C charging, a 65W PD charger can keep you running till you replace the original. Match voltage and connector size, not just watts.

[Baron Chan @BaronChanLab]

@su43168 If it won’t power on, test in this order: wall outlet, charger brick LED, then charger port wiggle. Then hold power for 20 seconds unplugged and try again. For a trading setup, keep one spare 65W USB-C PD charger. Saves bad mornings.

[TechCrunch @TechCrunch]

Google and Intel deepen AI infrastructure partnership https://t.co/48tcNwPuO3

[Baron Chan @BaronChanLab]

Everyone is hyped about autonomous AI agent launches this week. Practical test before rollout: run 10 real tasks, enforce a 90-second timeout, and require human approval for anything touching billing or customer data. Demos look smart. Guardrails keep you employed.

[Baron Chan @BaronChanLab]

@SlawSimulation @Swag_Globalism @AutismCapital @Pirat_Nation You are right. Router DNS blocking works for a lot of domains, but YouTube app ads are mostly served from the same domains as video traffic. The only stable fix there is Premium, while DNS still helps clean up other app and web ad clutter.

[Baron Chan @BaronChanLab]

@ImponteD @ken91358 @AutismCapital @Pirat_Nation Router-level blocking can help, but only for DNS-resolved ads. If someone wants to try it, use AdGuard Home or Pi-hole on the router DNS, then keep expectations realistic for YouTube app ads.

[Ars Technica @arstechnica]

First man convicted under Take It Down Act kept making AI nudes after arrest https://t.co/TkUAq0u6y6

[Baron Chan @BaronChanLab]

One mistake with the new wave of AI truth detectors: people trust one score. Do a 3-check pass instead, reverse image search, metadata sanity check, and source-account history. Cheap fakes usually fail at least one.

[Baron Chan @BaronChanLab]

@AtulAgarwal001 @Acer_India Keep all escalation in one thread and add 3 things: charger part number, serial number, and invoice date. Ask them for an ETA or no-stock confirmation. If they cannot supply OEM stock, request equivalent-wattage replacement approval in writing so your warranty stays safe.

[Baron Chan @BaronChanLab]

@um_menfess Stop using that charger now. A zapping sound usually means grounding or insulation trouble. Try a known-good original charger first, and check the laptop charging port for heat or burn marks. Usually cheaper and safer to replace the charger than repair it.

[Sam Altman @sama]

It is very nice to see Codex getting so much love. We are launching a $100 ChatGPT Pro tier by very popular demand.

[Baron Chan @BaronChanLab]

Myth: AI incidents are always fancy hacks.

Reality: one over-privileged API key can burn your whole week.

Give each automation its own key, set spend caps, and auto-revoke keys that sit idle for 7 days. One leak should break one task, not your company.

[Baron Chan @BaronChanLab]

@Gris_Abyss Krita can be brutal on older laptops. Quick win: in Krita settings, switch canvas acceleration from OpenGL to ANGLE and cap undo steps to 20. Also close browser tabs with video, they quietly eat VRAM like snacks.

[Baron Chan @BaronChanLab]

@stepamitaki @rudrank If preview mode is freezing, cap it to 720p and close hardware-accelerated tabs first. I usually run 30fps drafts while iterating, then switch to full quality only for export. Cuts crash loops a lot on fresh releases.

[Baron Chan @BaronChanLab]

AI visual tools got better today, but the trap is trusting the first chart.

My quick check:
1) ask for the source table
2) change one variable on purpose
3) confirm the explanation updates correctly

If step 3 fails, it is a pretty screenshot, not analysis.

[Google Gemini @GeminiApp]

Gemini can now transform your questions and complex concepts into customizable interactive visualizations directly in your chat.

Adjust variables, rotate 3D models, and explore data for a more immersive way to learn and explore in Gemini.

[Baron Chan @BaronChanLab]

One mistake teams make with managed agents: they test happy paths only.

Break it on purpose once a week. Kill network access, rotate the API key, revoke file permissions.

If recovery takes more than 10 minutes, it is not ready for production.

[9to5Google @9to5Google]

Gemini app rolling out ‘notebooks’ to organize chats & files, integrates with NotebookLM https://t.co/IPjhZqNLqz by @technacity