Ryan Mann

839 posts

Ryan Mann banner
Ryan Mann

Ryan Mann

@ryancmann

Analyst/modeler focused on DERs & demand flexibility. Technology + econ + regulatory/policy. @Sunrun @wattTime @Energy_Leaders @DER_Task_Force @theclimatevote

San Francisco Bay Area شامل ہوئے Haziran 2011
239 فالونگ263 فالوورز
Ryan Mann
Ryan Mann@ryancmann·
@ArushiSF And notably, many/most of those edge inference devices (smartphones, laptops) already have built-in batteries, so they are not just efficient consumers of power but also time-flexible ones.
English
1
0
1
16
Arushi Sharma Frank
Arushi Sharma Frank@ArushiSF·
"...AI stack will bifurcate.Heavy training & large-scale model hosting will remain centralized.But the bns of everyday inferences that define UX, such as voice recognition, image enhancement, industrial monitoring,or predictive maintenance, will[*] occur on the edge."
English
1
0
0
90
Arushi Sharma Frank
Arushi Sharma Frank@ArushiSF·
"@Apple, Qualcomm, AMD, & @intel are all embedding neural processing units directly into consumer hardware. Efficient model architectures, such as quantization, pruning, & distilled small language models, now allow impressive inference to happen on chips drawing a few watts." 🧵
English
5
1
6
1.9K
Ryan Mann
Ryan Mann@ryancmann·
@sdamico @ImpulseLabs_ That's pretty cool! I guess that makes sense that if the stove electronics can measure the changes in magnetic field from rotating the knob that you can also measure the change from pressing it.
English
0
0
0
12
Ryan Mann
Ryan Mann@ryancmann·
@nathan2grid The energy efficiency world has talked about this for a while - there's something called the Ratepayer Impact Measure test that looks at rate impacts on non-EE-program-participants. Most EE would fail the test, so it's not used much anymore.
English
0
0
0
75
Ryan Mann
Ryan Mann@ryancmann·
@sdamico You have temperature sensors inside the cooktop, and then use a state-estimation model (ex. KF) to estimate the temp of the top of the pan? Some of @ScottMoura's former students whose work we built upon did something similar w/ a fridge & its contents.
English
0
0
0
119
Ryan Mann ری ٹویٹ کیا
Adam Forni
Adam Forni@xternality·
#EnergyTwitter SF happy hour at Gotts (Ferry Building) 5pm next Monday 14th, reply or DM if attending. Some of us Google Climate team folks in town, and though I don't live in SF anymore, I sure enjoy a good meetup.
English
3
1
5
710
Ruaridh Macdonald
Ruaridh Macdonald@rrmacd·
For anyone modelling long-duration storage in capacity expansion models, Federico Parolin - a visiting PhD student @mitenergy, has just published a new formulation which reduces runtimes by at least 30% in our tests and avoids common constraint violations
English
1
6
18
3.1K
Ryan Mann
Ryan Mann@ryancmann·
@NiyerClimate Posit makes it easy and free to deploy data-vis web apps on shinyapps.io - used that for a CA Net Billing Tariff dashboard last year. Not sure if there's equivalent for Dash & Streamlit.
English
0
0
1
36
Ryan Mann
Ryan Mann@ryancmann·
@rohanspatel @wmorrill3 Have thought for a while that a version of Conway’s Law applies to the electric grid & the orgs that govern & manage it. The impact of distributed energy resources seems to currently be limited by the awkward abstractions used to integrate them into centralized planning/control.
English
0
0
2
31
Rohan Patel
Rohan Patel@rohanspatel·
This is such an awesome post from @wmorrill3 and applies to every organization. Please read it.
Wes@wmorrill3

Have you heard of Conway's Law? It's an interesting observation about the root cause of why large organizations usually make products worse. In 1967 Melvin Conway wrote "Organizations which design systems are constrained to produce designs which are copies of the communication structures of these organizations." When you think about any product carefully, if you see an obvious lack of integration it's usually due to an organizational social boundary. If you are working on a product you should consider if organizational boundaries are getting in the way of optimal design. A well known example - early days of Tesla there was a battery team and separately a vehicle structures team. Structures team designed their vehicle body to meet given requirements of strength, crashworthyness, torsional stiffness, etc. likewise the battery team designed their part to be self contained, it could survive durability, accidentally being dropped, being hit in a crash, etc. As a result, we ended up with was a super dense battery in a strong box like structure, which was then Installed into the vehicle which had a nice space for it to mate into. There were no issues with integration, everything fit together perfectly and met all product goals. It achieved one of the highest crash safety ratings measured at the time. But we had a box full of battery cells that was installed into another empty box shaped receptacle on the body. A box in a box. When you simplify it down that far it sounds obviously wrong. The two organizations had achieved their goals, worked together without friction, and the product met its overall goals. Yet the product ended up with a clear lack of optimization as a result of the organizational boundaries of the two teams working in isolation. Nothing was wrong, but it wasn't optimal. So before the next product was designed, the battery team gave responsibility of the battery structures also to the vehicle structures team. On this iteration, we ended up with the structural battery, which is an integral part of the body and crash structure. Without it, the vehicle body will not work. It's the literal floor for the vehicle. But the redundancy is gone and the design is more efficient as a result. This vehicle also achieved one of the highest crash safety ratings measured at the time. This is a super obvious example (in retrospect) and solved with a fairly large organizational change but you can also see this happen in small technical decisions and doesn't require structural change to fix. Someone just needs to question if there is a better solution in a team open to criticism. This mindset to work together to make the best product regardless of ego is where you end up with the most innovative products. Some smaller examples have been seen when inspecting Cybertruck design. The chassis air suspension which is used to pressurize the battery pack to prevent water ingress. The subwoofer which utilizes the air volume of the body side instead of making the enclosure larger. Centralized zonal vehicle controllers instead of many small distributed controllers. Doors which use the exterior surface as a crash intrusion beam. The pedestrian warning system used as a horn. The list goes on. The excitement and motivation by everyone involved to work across boundaries and actively break down Conway's Law is one of the many reasons I love working at Tesla.

English
5
5
211
13.7K
Ryan Mann
Ryan Mann@ryancmann·
@curious_founder Reminds me of this @daraobriain comedy bit about a time-traveler who can't explain how electric appliances work beyond "it's connected to the wall." youtube.com/watch?v=yJ0T6E… My Power Systems Engineering professor Sascha von Meier had this as the very first slide on Day 1:
YouTube video
YouTube
Ryan Mann tweet media
English
0
0
3
30
Daniel
Daniel@dkyost·
@c_r_spears @Climate_AF Is there a link of state races that are high impact we should be giving to now? Especially ones that are in swing states there may be some reverse coattails?
English
2
0
1
58
Ryan Mann
Ryan Mann@ryancmann·
@adambrowning I have accounts on all 3: Bsky feels the most like Twitter. Less activity, but has potential. Masto is even more quiet, usability isn't great, have heard the active-user culture is a bit purist/exclusionary. Threads feels like more Instagram, not much climate/energy traction.
English
0
0
1
33
derek guy
derek guy@dieworkwear·
The only objectively bad knots are these, which tell people you learned how to tie your tie on Reddit. Please don't wear stuff like this.
derek guy tweet media
English
105
135
3.8K
229.8K
derek guy
derek guy@dieworkwear·
A few people asked me about this, so I will do a thread. As usual, the story starts with a bit of history about socio-economic class. 🧵
English
124
525
8.6K
2.3M
Ryan Mann
Ryan Mann@ryancmann·
@ArushiSF The @Energy_Leaders fellowship includes a session on retail ratemaking, and the fellows include a lot of people working for innovative orgs. Could try to connect you with one of the fellowship coordinators if you'd be interested in being a speaker.
English
2
0
1
30
Arushi Sharma Frank
Arushi Sharma Frank@ArushiSF·
If you thought I wasn't being serious about this I was. Here's proof. How many clean energy companies startups innovators do you see on this list? It's a public registration list for a highly reputable rate making course. cpu.nmsu.edu/conferences/re…
Arushi Sharma Frank@ArushiSF

6. Too many people who work in the clean energy transition do not understand fundamentals of cost of service rate making: the number one job of PUCs is economic reg. open.substack.com/pub/arushishar…

English
2
0
1
236
Ryan Mann
Ryan Mann@ryancmann·
@NiyerClimate What plotting package is that? And are you using a software development environment with an LLM assistant like GitHub Copilot built in, or an external chatbot?
English
0
0
2
61
Ryan Mann
Ryan Mann@ryancmann·
@posamentier My understanding is that they're conducting a study of 4 different alternative propulsion techs + diesel baseline so that other transit agencies can learn from their experience. Also rooting for the battery one! mobility-innovators.com/ac-transit-zer…
English
0
0
0
96
Kyle Baranko
Kyle Baranko@kyle__cb·
Flexible interconnection trust issues 1. Utility doesn’t trust large load / new gen to behave according to time varying local constraints so they oversize the system upgrade or lock you into artificially low consumption / output with static schedule, OR insist on operating the asset 2. Developer doesn’t trust the utility to operate the asset for fear of leaving value on the table / operating sub-optimally and providing bad end customer experience Solution is a smart contract with mutually-agreed upon load shapes that are dependent on a verified source of input data (on oracle, if you will) that both parties have access to This data feed would be near real time SCADS readings of the equipment causing the time varying local constraint Equipment modulating the power flow at the POI responds to the terms of the contract and the data feed, not either party’s in house system Result is that the utility does not have the ability to shut off power unless conditions of contract are met but customer has no ability to pull a larger load of inject more power than what they have previously agreed to Software for this exists, but does the hardware?
English
9
4
24
2.7K