Shyam D. PhD
@PeakSquirrel
Followers
788
Following
248
Media
46
Statuses
2,655
Mostly human, but a bit squirrel. Still can’t locate his acorns. My own views. He/Him
Joined August 2019
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
#Comebacktome
• 612235 Tweets
#RightPlaceWrongPerson
• 390502 Tweets
RM IS COMING
• 317962 Tweets
namjoon
• 265030 Tweets
#LoseMyBreath
• 231941 Tweets
Official MV
• 144091 Tweets
メイドの日
• 113708 Tweets
Luka
• 83173 Tweets
Celtics
• 79847 Tweets
भगवान परशुराम
• 75804 Tweets
#ดวงใจเทวพรหมSportsDay
• 70197 Tweets
メイドさん
• 50888 Tweets
Megan
• 42184 Tweets
メイド服
• 40010 Tweets
भगवान विष्णु
• 38760 Tweets
ドラマ化
• 30620 Tweets
Mavs
• 29235 Tweets
Rayados
• 22129 Tweets
RTTT
• 18074 Tweets
Sosa
• 17996 Tweets
Tigres
• 17879 Tweets
太陽フレア
• 15286 Tweets
Canales
• 15208 Tweets
Pumas
• 14792 Tweets
Hayırlı Cumalar
• 12197 Tweets
Samir
• 11789 Tweets
Chet
• 11311 Tweets
#AkshayaTritiya
• 11192 Tweets
IS:SUE
• 10193 Tweets
I can't emphasize this enough. In Industry, I can just change companies if my environment turns toxic. In Academia, I had to quit my speciality, because no matter where I went, its the same "community" of peers controlling my access to a job, funding, awards, etc.
3. I wanted to be treated with respect.
In academia, especially as a post-doc, I didn't feel valued. I was actually bullied by a tenured prof in my dept. Unfortunately, my experience is not unique.
Now, my colleagues often tell me how much my hard work is valued. 4/
2
5
162
7
17
195
After 5.5 years at
@BerkeleyLab
, I'll be moving to
@LIST_Luxembourg
to start a brand new group going head to head with the likes of
@A_Aspuru_Guzik
,
@aicooper
,
@leecronin
,
@brycemeredig
, and
@cedergroup
to accelerate the synthesis of novel materials.
14
1
136
Oh boy. This is a great thread to unpack how the experimental world is being held back when thinking about the role of AI/ML. A unfortunately long thread
This exciting paper shows AI design of materials, robotic synthesis. 10s of new compounds in 17 days.
But did they?
This paper has very serious problems in materials characterisation. In my view it should never have got near publication. Hold on tight let's take a look 😱
64
514
3K
8
6
103
@clairevo
Everything you're saying sounds like managing up. That might be the practical solution for an individual. It should never be the solution for an organization. Thats just mismanagement. It's not the individual engineers job to do the PM, Director, VP, etc.'s job.
1
0
87
The annual Materials Project Workshop will be virtual again this year from August 10 - 12. Sign up for 3 days of materials informatics training. We'll even teach you to code in python!
#Chemistry
#CompChem
#Informatics
1
17
63
@ashleyruba_phd
@Nature
I'm not going to be as harsh on post-doc positions if you understand and use them in the way the current system treats them: A temporary position to bridge you to where you want to go. That means you primary job is to figure out what you're doing next. Not research.
6
0
60
@PhTaheri
Disagree. Teaching/Mentoring are from the same skillset IMO. I think we need a system that values teaching enough to provide the resources: space, time, positions, salary, and prestige to those that do well at teaching; equal to our best researchers.
1
0
55
@academic_exit
3.5X postdoc salary and I had a high postdoc salary thanks to the national labs
0
1
55
I'll give a good example of how complicated academic budgeting for just postdoc labor can be; in this case at Lawrence Berkeley Labs.
#AcademicChatter
#Budgets
6
11
47
@J_A_C_S
Let's start with an even simpler yet more asinine problem. All data in figures must be in raw form in the SI. In 2021, we shouldn't have to extract data from PNGs and JPEGs.
3
2
43
The Materials Project Workshop is going virtual this summer. Join us for 2+ days of highly interactive training on the open-source software tools for data-driven materials science developed by the Materials Project.
0
12
35
102 Virtual Participants / 3 Days / 21 Hours
Our first virtual Materials Project Workshop was a resounding success.
Thanks to everyone from
@persson_group
that helped make this happen!
#python
#compchem
#workshops
#MPWorkshop2020
0
11
31
@pauldauenhauer
I love how people are nit picking your numbers. They do understand that you have almost 0 control over costs? They get decided by VP and Deans and then forced all the PIs who have to write grants to bring this amount of money to the university.
1
0
31
@martinmbauer
@heinrichhartman
Journal based Peer Review should be abolished. We do need a system of review, but the absolute molotov-flaming-dump-truck of our current system needs to go. That won't happen by finding the right one. It will happen by trying different options and be willing to fail along the way
1
0
27
@JohnStreicher1
There's no question that a PhD increases your earning potential, but outside of academia many of those positions can be obtained without a PhD and while doing jobs that pay far higher than PhD stipend. Is a PhD still worth it then?
5
0
29
@shortstein
I'd argue the problem is not the number of papers, but the laughbly low standards for authorship for senior researchers vs early career reseachers. ECRs have to do work to be an author, but many senior PIs demand authorship for simply being the boss of an ECR who contributed.
3
1
29
Great MP Seminar by
@lonepair
. Only he could confirm that all of us high-throughput computational chemists are just trying to climb out of the depths of data-hell.
#chemtwitter
#MaterialsProject
1
1
26
@micahgallen
I don't think we even need Journals. Fully open and community directed peer review on
@PubPeer
is far superior to the highly secretive peer review process we have now. Just look at the fraud
@MicrobiomDigest
finds as a single individual.
5
6
24
@rjurney
@CollinMickels
I think that's the point. Those highly capable engineers make 2x-3x just by switching to ML or SWE, and they're very adept at it so the market supports draining the entire experimental workforce if it went that far.
1
0
23
@georgemporter
The cloud takes staff to use well. Labs have lab managers. Funders don’t understand this and won’t fund the staff. I wonder if we need to build an academic cloud just for these kinds of edge cases.
2
0
24
since
#postdoc
are apparently traineeships, it would be nice if postdoc postings mentioned what skills the trainee would gain instead of just "job requirements"
2
0
23
@Kyle_Ireton
I'll go:
Grad Student in Michigan: $26K (2010 - 2016)
High Salaried PostDoc at LBL: $70K (2016 - 2018)
Staff Scientist at LBL: $110K (2018-2021)
Staff Data Engineer in the SF Bay area, TC: $300K ($200K Salary)
4
2
20
@CollinMickels
The other side of this coin is that these companies have no problem hiring a SWE for SWE salaries. They make it very clear what skills they value to the very people they are losing.
0
0
20
@Robotbeat
Can you point to which Chinese labs have completed their attempts and put up results? I'm not seeing any.
3
0
21
@njoybetz
My experience has been that none of us knew what we were getting into in applying for a PhD, PostDoc, or even PI-position at a National Lab or University (Professor). It's not advertised or taught anywhere.
1
0
19
@simonw
Figuring out which problems are AI-ready, how to convert them to an AI problem, and then deploying whatever works is a massive problem few people seem to focus on.
1
0
19
What's the point in making the most widely used database of materials data WITH a FAIR API if reviewers are still going to complain that the data is not available as giant intangible zip file on Figshare?
#AcademicTwitter
#chemtwitter
#FAIRdata
#MaterialsProject
2
1
18
@JohnDiesattheEn
Hey
@Google
, I think you have a problem. These aren't the desk ornaments we're searching for.
0
1
16
Annual Materials Project Workshop scheduled for July 28th - July 30th. Register now for a spot!
#MaterialDesign
#Python
#DataScience
@KPatBerkeley
@persson_group
0
16
18
@PhD_Genie
Communication - actually getting the other person to understand. Collaboration- actually working as part of a team.
0
1
16
One step closer to understanding how disorder affects complex properties
@persson_group
@KPatBerkeley
@BerkeleyLab
Origin of Disorder Tolerance in Piezoelectric Materials and Design of Polar Systems
0
7
15
@Juliamorphology
Suspect
#3
like everyone else. Diversity quotas should be in actual hires not the application pool. Otherwise, they just end up hurting minorities more because of situations such as this.
2
0
17
@jenheemstra
@cenmag
Changing the fundamental incentives: grants, honors, positions, and tenure to actually prioritize diversity, inclusion, mentorship, and collaboration.
0
0
17
@TatemRios
Grad school is all about learning. Switch and learn. For ref, I did an experimental PhD and then went into DFT and informatics in my postdoc + staff career. My exp background meant i could interface with a lot more collaborators with less effort to "earn" their trust.
1
0
17
If postdocs are trainees, why are they expected to come in already having all the skills necessary for a project?
2
1
17
@drrachelbrenner
The fact that a huge portion of academic Twitter doesn’t see it as satire might indicate that it fails at satire. The reality is that a lot of those statements are well within the norm for academia. That’s not satire.
0
0
15
@rita_strack
You gotta be specific with the genie. We asked for this in engineering and got two decades of matlab. Only now is it becoming python or Java .
2
1
15
@lilopozzo
I love this idea. Students compete for funding from NSF and then PIs compete for students. This changes the power dynamic in a good way.
1
1
15
I learned all of this at the drivers wheel. I had to be responsible for a budget before I was given access to the documentation + the Resource Adminstrators who knew how to interpret the rules to understand all of this. No Postdoc could get this info at LBL.
3
1
15
@CT_Bergstrom
Compare this with the notification
@stripe
and
@patrickc
sent out about their workforce reduction. Speaks volumes about the quality of leadership.
1
0
15
@MaximZiatdinov
Because the people who have enough understanding of both get gobbled up by industry with $300k+ compensation plans
2
0
14
@AnnaGHughes
I call your Shrek 2 and raise you one Michael Bay movie cuz the explosions keep me awake for the overly expensive food cart I get 20% off on with my airline CC.
#Thrifty
0
0
14
@bernhardsson
Hmmm, i'd argue for version control. Pretty sure all software would be 10,000 lines of uncommented to every 1 line of functioning code without it.
1
1
13
@vikasnavaratna
I would agree with this sentiment if students and postdocs got even a half-unbiased view of the world. Industry is almost always portrayed as lesser and more evil in academia, when in reality they are equal levels of good/bad.
2
0
14
@MCDuncanLab
This is brilliant and should be the core aim of a STEM PhD. I don't agree that people can just figure this out and the fact that you did shows amazing self-awareness.
2
0
13
I’ve been learning to use
@DataPolars
as a MUCH faster alternative to Pandas and stumbled upon this: the derived_from syntax alone is very enticing. Add
@pydantic
compatibility = Wow.
@RitchieVink
1
3
13
@shunley42
What you actually need to know:
- How to talk to your customers and figure out WTH they actually want
- How to say no
- How to keep track of what you did and why (Aka Version Control)
- How to do the above as a part of a team
- Pick One: SQL, Python, R, PowerBI, Tableau, etc.
1
0
13
@ashleyruba_phd
I'd go further and suggest we have to swap the power dynamic. Grant agencies must fund students and postdocs directly. The funds then follow them.
1
0
13
@Kyle_Ireton
@academic_exit
This. We're not forced to change. We have the option to change to improve w.e. we want: pay, culture, curiosity. If anything the freedom can be greater than in academia where you are stuck in the funding you have/can get.
0
0
13
@vikasnavaratna
I'm honestly amazed at how much grocery shopping you and your wife must do to support the stream of diverse and amazing food pictures.
1
0
13
@ramav_matsci
Go straight for Global Entry which gives you TSA Precheck as well. GE is worth its weight in gold for getting me out of slow customs lines.
1
0
11
@Melterialsgal
Agreed on the academic side. On the industry side they still feel like flashy metrics. Pretty sure Goodharts law is playing a strong role in all CVs.
0
0
10
@spencer_jolly
@chrisdc77
Oh! No! It could allow a few bad apples. I'd much rather keep what we've got. *looks at the rotting pile of horse manure that is academic merit, hiring and promotion*
0
0
12
I decided to jam out to 2000's rock today. It was such a good decade. Linkin Park, Evanescence, Foo Fighters, White Stripes, 3 Doors Down, Stone Sour, System of a Down, Disturbed and so many more were next level.
5
1
12
@jmdagdelen
attracting the largest crowd of anyone at his poster on NLP for Materials Design.
@persson_group
#Coda2020
0
2
11
I just saw an intern ad at Microsoft with a higher salary then my previous Scientist position at a National Lab
#AltAc
#AcademicChatter
2
0
12
1
4
11
It seems like most
#deeptech
startups in energy are a spin out of an
#academia
R&D project. Any suggestions on how to run that fundamental R&D outside of academia? Are there funding/labs that let individuals try out small scale experiments?
@AltAcChats
0
3
12
@bernhardsson
They have it in the recipes package more-itertools. Still don’t know why it’s a separate thing.
1
0
11
At a technical level, it's not clear most people responding read the paper. Closing the loop and fast discovery via AI/ML was the point. That necessitates dealing with low-quality measurements and exploratory decision-making.
2
1
12
@njoybetz
@AltAcChats
I think this is a myopic view. Don’t forget that nearly every Ivy League endowment is full of money from slavery. Universities have become corporations since Bayh-Dole. Even scientific societies like ACS act like a corporation. They all want money.
1
0
10
Night 1 of international move. Already have really bad imposter syndrome. It’s like I’ve been training this reaction my whole life.
#mentalhealth
#movingtodifferentworld
#AcademicMentalHealth
1
0
11
@likethebuilder
Honestly doesn’t surprise me at all. We all have to decide when we stop going down the rabbit hole to work on our plane. Some do a better job of that decision.
0
0
11
@Kyle_Ireton
I'm going to bet that said PI does none of the "training" and it rather comes from the rest of the group of "trainees".
0
1
11
@mike_feigin
Communication training. Specifically the workshops by Jean-Luc Dumont completely changed how I write, make presentations, document, etc.
0
0
11
This was my salary as staff scientist at LBL and > double what UC Berkeley pays for a postdoc. Not to mention working on cutting edge work. As a former Nuclear Engineer I would have been stoked to have this opportunity.
Prestigious + highly paid ($110k salary)
#postdoc
opportunity
@BerkeleyLab
for US citizens - come work with me at the intersection of nuclear chemistry,
#compchem
, reaction networks, and
#MachineLearning
! RTs appreciated!
#ChemPostdoc
More info:
1
10
28
0
1
11
@darren_lipomi
I honestly think funding sources should include lifetime caps adjusted for overhead. There's no way that a PI with 30 funded postdocs can adequately manage/mentor them to be intellectually included on all those papers. It's insane how few PIs drain most of the funding.
0
0
11
@ChemicalScience
@SamMBlau
@hpatel1567
@EWCSpotteSmith
@KPatBerkeley
Check out our new paper on building reaction networks that preserve stoichiometry during pathfinding. Congrats on
@SamMBlau
for doing such an amazing job on such a cross-disciplinary topic!
@persson_group
@BerkeleyLab
0
0
11
I think of the move from traditional algorithms to ML/AI as a switch from deterministic guarantees to probabilistic guarantees. This is akin to the move from thermo to stat mech; from newtonian to quantum mechanics. This makes sense in my head.
1
0
11
@phylogenomics
@NobelPrize
One is an “engineering” problem, the other is “science”. It’s a highly flawed system that rewards serendipitous science over purpose-driven problem solving.
0
0
11
@olexandr
@AmerChemSociety
@RoySocChem
You mean the ACS that pays its CEO $1M+ per year by charging thousands of dollars in registration fees and publication fees? MDPI is bad, but let's not pretend that publishers are anything but shades of grey bad these days.
1
0
10
Super unpopular take.
PhDs should have to write a proper thesis. Not N papers stapled together, but a coherent thesis. This is how PhDs learn a proper scientific method at the scale of modern science. It's also the most general and transferable skill Academia has to offer:
5
1
10
Is it just me or does the software defined assets in
@dagsterio
do much of acryl data hub and Atlan’s offering, but like 1/1000 the bloat? Why do we need a SQL db, Kafka and elastic search for metadata and lineage?
0
1
10
@seis_matters
I'm a firm believer that "Insert problem in Academia" somehow ties back the basic issue that Publications are a metric for anything other than "Number of Publications"
0
0
10
@njoybetz
Fidelity has good rates for their savings account and their automatically CD-ladder is a fantastic product. Have CDs expire at fixed intervals and get reinvested into a new CD at the best market rate.
0
0
10
The most important thing Industry gave me was a way to disconnect my identity from my job. I am not my job. So changing jobs every year to get more pay or a better role isn't an issue. I don't have to "wait for my turn" or tenure or someone else to finally retire.
Perhaps some of our trainees think about leaving academia for the wrong reasons. Industry will not give you more job security! Despite all the negative stories on this platform, I think academia can provide a highly fulfilling career, and we should communicate this to them.
36
21
196
1
2
10
@scanlond81
Sounds nonsensical. Why, as a PI, would you want to turn away students that might have had valuable experimental experience and skills to provide context and benchmarks?
1
0
10
Any DevOps engineers tired of building another social media platform? Come join my team at
@Rivian
and build scaleable infrastructure for battery manufacturing. Make an impact on the future of transportation, energy and climate!
0
1
9
@XtalGrrl
I blame the instrument makers who have produced "good enough" tools that seem easy to use but in fact hide a lot of the complexities. A few decades of this and "anyone" can analyze a PXRD pattern.
2
0
9
I will admit. Job hunting sucks, even when it’s going well.
1
1
9
@AltAcChats
Staff Data Engineer supporting the next generation of battery manufacturing for EVs
@Rivian
0
0
9
@Kyle_Ireton
I’m glad this worked out for you, but it would very problematic for someone that is interested in companies closer to gov contracts, where the change in title could cause legal consequences. Just have to mindful.
2
0
9
@MaximZiatdinov
I'm there now too. A bit sick and tired of "AI accelerated XYZ" with no plan to actually solve any relevant problems.
0
0
9
@adamcbest
Maybe we need a new branch of the Armed Forces: "Climate Force" or "Special Weather Operations"? That way we can direct that$840B to just that.
0
2
9
Listening to "The Making of the Atomic Bomb" by Richar d Rhodes and amazed by how much of Nobel-winning research was: "Irradiate stuff and watch cloud chamber". None of this stuff would ever get "funded" today.
#AcademicTwitter
#GoodhartsLaw
2
0
9
Had a meeting with a clearly defined goal and it was great. Now I understand I don’t hate meetings, just bad meetings which are the majority.
0
0
7
@dkjhaunc
@TrumanLab
Knowing who foots the bill and the details of that bill are two different things. Even then there are so many "secrets" aspects of that bill that are not discussed with "trainees" even though thats probably the most important factor in getting that funding.
1
0
8
Changing your password every X days is like daylight savings. It might help someone, but it hurts a whole lot of other people collectively way more.
#why
#passwords
#CyberSecurity
2
0
9
I've learned so much from the numerous wonderful friends in the
@persson_group
.
@KPatBerkeley
continues to set the standard for an open, inclusive, compassionate, and team-oriented environment that I have yet to see anywhere else in academia.
1
0
9
@MolecularXtal
This is honestly a huge problem in
#MaterialsProject
and
@pymatgen
where even core contributors got 0 credit for the immense amount of work we put in because we weren’t on the original paper. Conversely people who did minimal for that initial paper are making citation bank.
2
0
9
One of the challenges for
#OSS
in the
#academic
community such as
#chemtwitter
is the inability to get credit for contributions. As a result every group seems to have their own (Chemistry, Materials) informatics package to claim ownership; most of which are just bad or copies.
2
1
9